cases/industrial_examples/energy_monitoring/building_energy_consumption.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "source": [
    "## Predict how much energy will a building consume with Fedot.Industrial"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Dataset published on Kaggle, aims to assess the value of energy efficiency improvements. For that purpose, four types of sources are identified: electricity, chilled water, steam and hot\n",
    "water. The goal is to estimate the **energy consumption in kWh**. Dimensions correspond to the air temperature, dew temperature, wind direction and wind speed. These values were taken hourly during a week, and the output is the meter reading of the four aforementioned sources. In this way, was created four datasets:  **ChilledWaterPredictor**, **ElectricityPredictor**, **HotwaterPredictor**, and **SteamPredictor**.\n",
    "Link to the dataset - https://www.kaggle.com/code/fatmanuranl/ashrae-energy-prediction2"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-08-28T10:34:48.354623Z",
     "start_time": "2023-08-28T10:34:39.594404Z"
    }
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "from fedot_ind.api.utils.path_lib import PROJECT_PATH\n",
    "from fedot.core.pipelines.pipeline_builder import PipelineBuilder\n",
    "from fedot_ind.tools.loader import DataLoader\n",
    "from fedot_ind.core.repository.initializer_industrial_models import IndustrialModels\n",
    "from examples.example_utils import init_input_data, calculate_regression_metric\n",
    "from golem.core.tuning.sequential import SequentialTuner\n",
    "from fedot.core.pipelines.tuning.tuner_builder import TunerBuilder\n",
    "from fedot.core.repository.quality_metrics_repository import RegressionMetricsEnum"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The list of basic fedot industrial models for experiment are shown below. We using simple linear machine learning pipelines with 3 different feature generators: Statistical, Reccurence, Topological. And for each of them we add PCA transformation with 90 % of explained dispersion."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "outputs": [],
   "source": [
    "model_dict = {\n",
    "    'regression_with_statistical_features': PipelineBuilder().add_node('quantile_extractor',\n",
    "                                                                       params={'window_size': 5}).add_node('ridge'),\n",
    "    'regression_pca_with_statistical_features': PipelineBuilder().add_node('quantile_extractor',\n",
    "                                                                           params={'window_size': 5}).\n",
    "    add_node('pca', params={'n_components': 0.9}).add_node('ridge'),\n",
    "    'regression_with_recurrence_features': PipelineBuilder().add_node('recurrence_extractor',\n",
    "                                                                           params={'window_size': 20}).add_node('ridge'),\n",
    "    'regression_pca_with_recurrence_features': PipelineBuilder().add_node('recurrence_extractor',\n",
    "                                                                           params={'window_size': 20}).\n",
    "    add_node('pca', params={'n_components': 0.9}).add_node('ridge'),\n",
    "    'regression_with_topological_features': PipelineBuilder().add_node('topological_extractor',\n",
    "                                                                           params={'window_size': 20}).\n",
    "    add_node('pca', params={'n_components': 0.9}).add_node('ridge'),\n",
    "    'regression_pca_with_topological_features': PipelineBuilder().add_node('topological_extractor',\n",
    "                                                                           params={'window_size': 20}).\n",
    "    add_node('pca', params={'n_components': 0.9}).add_node('ridge')\n",
    "}\n",
    "\n",
    "dataset_name = 'ElectricityPredictor'\n",
    "data_path = PROJECT_PATH + '/examples/data'\n",
    "tuning_params = {'task': 'regression',\n",
    "                 'metric': RegressionMetricsEnum.RMSE,\n",
    "                 'tuning_timeout':10,\n",
    "                 'tuning_iterations':30}"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "outputs": [],
   "source": [
    "from functools import partial\n",
    "\n",
    "\n",
    "def evaluate_industrial_model(train_data, test_data, task:str = 'regression'):\n",
    "    metric_dict = {}\n",
    "    input_data = init_input_data(train_data[0], train_data[1], task=task)\n",
    "    val_data = init_input_data(test_data[0], test_data[1], task=task)\n",
    "    with IndustrialModels():\n",
    "        for model in model_dict.keys():\n",
    "            print(f'Current_model - {model}')\n",
    "            pipeline = model_dict[model].build()\n",
    "            pipeline.fit(input_data)\n",
    "            features = pipeline.predict(val_data).predict\n",
    "            metric = calculate_regression_metric(test_target=test_data[1], labels=features)\n",
    "            metric_dict.update({model: metric})\n",
    "    return metric_dict\n",
    "\n",
    "def tuning_industrial_pipelines(pipeline, tuning_params, train_data):\n",
    "    input_data = init_input_data(train_data[0], train_data[1], task=tuning_params['task'])\n",
    "    pipeline_tuner = TunerBuilder(input_data.task) \\\n",
    "        .with_tuner(partial(SequentialTuner, inverse_node_order=True)) \\\n",
    "        .with_metric(tuning_params['metric']) \\\n",
    "        .with_timeout(tuning_params['tuning_timeout']) \\\n",
    "        .with_iterations(tuning_params['tuning_iterations']) \\\n",
    "        .build(input_data)\n",
    "\n",
    "    pipeline = pipeline_tuner.tune(pipeline)\n",
    "    return pipeline"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now we must download the dataset. It could be done by using `DataReader` class that implemented as attribute of `FedotIndustrial` class. This class firstly tries to read the data from local project folder `data_path` and then if it is not possible, it downloads the data from the UCR/UEA archive. The data will be saved in the `data` folder."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-08-28T10:35:13.321212Z",
     "start_time": "2023-08-28T10:35:12.913025Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 15:42:50,001 - Reading data from D:\\WORK\\Repo\\Industiral\\IndustrialTS/examples/data/ElectricityPredictor\n"
     ]
    }
   ],
   "source": [
    "_, train_data, test_data = DataLoader(dataset_name=dataset_name).read_train_test_files(\n",
    "            dataset_name=dataset_name,\n",
    "            data_path=data_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "outputs": [],
   "source": [
    "input_data = init_input_data(train_data[0], train_data[1], task=tuning_params['task'])\n",
    "val_data = init_input_data(test_data[0], test_data[1], task=tuning_params['task'])"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Lets check our data."
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "outputs": [
    {
     "data": {
      "text/plain": "(567, 4, 168)"
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "input_data.features.shape"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Lets visualise our predictors."
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "outputs": [
    {
     "data": {
      "text/plain": "<Figure size 1100x400 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": "<Figure size 1100x400 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": "<Figure size 1100x400 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": "<Figure size 1100x400 with 1 Axes>",
      "image/png": "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"
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "pd.DataFrame(input_data.features[1, 0, :]).plot(title='air temperature')\n",
    "pd.DataFrame(input_data.features[1, 1, :]).plot(title='dew temperature')\n",
    "pd.DataFrame(input_data.features[1, 2, :]).plot(title='wind direction')\n",
    "pd.DataFrame(input_data.features[1, 3, :]).plot(title='wind speed')\n",
    "plt.show()"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "Next steps are quite straightforward. We need to fit the model and then predict the values for the test data just like for any other model in sklearn.\n",
    "\n",
    "At the `fit` stage FedotIndustrial will transform initial time series data into features dataframe and will train regression model."
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "start_time": "2023-08-28T10:35:27.965798Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Current_model - regression_with_statistical_features\n",
      "Current_model - regression_pca_with_statistical_features\n",
      "Current_model - regression_with_reccurence_features\n",
      "Current_model - regression_pca_with_reccurence_features\n",
      "Current_model - regression_with_topological_features\n",
      "Current_model - regression_pca_with_topological_features\n"
     ]
    }
   ],
   "source": [
    "metric_dict = evaluate_industrial_model(train_data,test_data)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "At the end of the experiment we can obtain the desired metric values using `calculate_regression_metric` method. Now there are five available metrics for classification task:\n",
    "- `explained_variance_score`\n",
    "- `max_error`\n",
    "- `mean_absolute_error`\n",
    "- `mean_squared_error`\n",
    "- `d2_absolute_error_score`.\n",
    "- `median_absolute_error`\n",
    "- `r2_score`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2023-08-28T11:01:34.941934Z",
     "start_time": "2023-08-28T11:01:34.928460Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": "                                          r2_score:  mean_squared_error:  \\\nregression_pca_with_reccurence_features    0.080487        242844.478589   \nregression_with_reccurence_features        0.038808        253851.946498   \nregression_with_topological_features       0.010349        261368.003356   \nregression_pca_with_topological_features   0.010349        261368.003356   \nregression_pca_with_statistical_features  -0.002798        264840.107063   \nregression_with_statistical_features      -1.644464        698406.082747   \n\n                                          root_mean_squared_error:  \\\nregression_pca_with_reccurence_features                 492.792531   \nregression_with_reccurence_features                     503.837222   \nregression_with_topological_features                    511.241629   \nregression_pca_with_topological_features                511.241629   \nregression_pca_with_statistical_features                514.626182   \nregression_with_statistical_features                    835.706936   \n\n                                          mean_absolute_error  \\\nregression_pca_with_reccurence_features            244.728677   \nregression_with_reccurence_features                258.780680   \nregression_with_topological_features               243.908111   \nregression_pca_with_topological_features           243.908111   \nregression_pca_with_statistical_features           261.164925   \nregression_with_statistical_features               604.766705   \n\n                                          median_absolute_error  \\\nregression_pca_with_reccurence_features              120.852159   \nregression_with_reccurence_features                  126.047284   \nregression_with_topological_features                 126.156833   \nregression_pca_with_topological_features             126.156833   \nregression_pca_with_statistical_features             133.701291   \nregression_with_statistical_features                 424.789912   \n\n                                          explained_variance_score  \\\nregression_pca_with_reccurence_features                   0.085936   \nregression_with_reccurence_features                       0.043402   \nregression_with_topological_features                      0.013599   \nregression_pca_with_topological_features                  0.013599   \nregression_pca_with_statistical_features                 -0.001518   \nregression_with_statistical_features                     -1.606454   \n\n                                            max_error  d2_absolute_error_score  \nregression_pca_with_reccurence_features   2167.000715                -0.268740  \nregression_with_reccurence_features       2183.389781                -0.341590  \nregression_with_topological_features      2345.116435                -0.264486  \nregression_pca_with_topological_features  2345.116435                -0.264486  \nregression_pca_with_statistical_features  2337.330582                -0.353950  \nregression_with_statistical_features      2413.925700                -2.135276  ",
      "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>r2_score:</th>\n      <th>mean_squared_error:</th>\n      <th>root_mean_squared_error:</th>\n      <th>mean_absolute_error</th>\n      <th>median_absolute_error</th>\n      <th>explained_variance_score</th>\n      <th>max_error</th>\n      <th>d2_absolute_error_score</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>regression_pca_with_reccurence_features</th>\n      <td>0.080487</td>\n      <td>242844.478589</td>\n      <td>492.792531</td>\n      <td>244.728677</td>\n      <td>120.852159</td>\n      <td>0.085936</td>\n      <td>2167.000715</td>\n      <td>-0.268740</td>\n    </tr>\n    <tr>\n      <th>regression_with_reccurence_features</th>\n      <td>0.038808</td>\n      <td>253851.946498</td>\n      <td>503.837222</td>\n      <td>258.780680</td>\n      <td>126.047284</td>\n      <td>0.043402</td>\n      <td>2183.389781</td>\n      <td>-0.341590</td>\n    </tr>\n    <tr>\n      <th>regression_with_topological_features</th>\n      <td>0.010349</td>\n      <td>261368.003356</td>\n      <td>511.241629</td>\n      <td>243.908111</td>\n      <td>126.156833</td>\n      <td>0.013599</td>\n      <td>2345.116435</td>\n      <td>-0.264486</td>\n    </tr>\n    <tr>\n      <th>regression_pca_with_topological_features</th>\n      <td>0.010349</td>\n      <td>261368.003356</td>\n      <td>511.241629</td>\n      <td>243.908111</td>\n      <td>126.156833</td>\n      <td>0.013599</td>\n      <td>2345.116435</td>\n      <td>-0.264486</td>\n    </tr>\n    <tr>\n      <th>regression_pca_with_statistical_features</th>\n      <td>-0.002798</td>\n      <td>264840.107063</td>\n      <td>514.626182</td>\n      <td>261.164925</td>\n      <td>133.701291</td>\n      <td>-0.001518</td>\n      <td>2337.330582</td>\n      <td>-0.353950</td>\n    </tr>\n    <tr>\n      <th>regression_with_statistical_features</th>\n      <td>-1.644464</td>\n      <td>698406.082747</td>\n      <td>835.706936</td>\n      <td>604.766705</td>\n      <td>424.789912</td>\n      <td>-1.606454</td>\n      <td>2413.925700</td>\n      <td>-2.135276</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_baseline = pd.concat([x for x in metric_dict.values()],axis=1)\n",
    "df_baseline.columns = list(metric_dict.keys())\n",
    "df_baseline = df_baseline.T\n",
    "df_baseline.sort_values(by='root_mean_squared_error:')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "outputs": [],
   "source": [
    "best_baseline = df_baseline.sort_values('root_mean_squared_error:')['root_mean_squared_error:'].iloc[0]"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Could it be done better? Tuning approach"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "outputs": [],
   "source": [
    "metric_dict = {}"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Current_model - regression_with_statistical_features\n",
      "2023-10-09 15:44:53,617 - DataSourceSplitter - Stratificated splitting of data is disabled.\n",
      "2023-10-09 15:44:53,618 - DataSourceSplitter - Hold out validation is applied.\n",
      "2023-10-09 15:44:53,620 - SequentialTuner - Hyperparameters optimization start: estimation of metric for initial graph\n",
      "2023-10-09 15:45:08,468 - SequentialTuner - Initial graph: {'depth': 2, 'length': 2, 'nodes': [ridge, quantile_extractor]}\n",
      "ridge - {}\n",
      "quantile_extractor - {'window_size': 5} \n",
      "Initial metric: [753.794]\n",
      "  0%|          | 0/15 [00:00<?, ?trial/s, best loss=?]2023-10-09 15:45:08,473 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:45:08,473 - TPE using 0 trials\n",
      "  7%|▋         | 1/15 [00:11<02:40, 11.49s/trial, best loss: 753.9496624585285]2023-10-09 15:45:19,966 - build_posterior_wrapper took 0.000975 seconds\n",
      "2023-10-09 15:45:19,967 - TPE using 1/1 trials with best loss 753.949662\n",
      " 13%|█▎        | 2/15 [00:23<02:34, 11.92s/trial, best loss: 753.9496624585285]2023-10-09 15:45:32,178 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:45:32,179 - TPE using 2/2 trials with best loss 753.949662\n",
      " 20%|██        | 3/15 [00:33<02:10, 10.85s/trial, best loss: 753.9496624585285]2023-10-09 15:45:41,759 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:45:41,760 - TPE using 3/3 trials with best loss 753.949662\n",
      " 27%|██▋       | 4/15 [00:44<01:59, 10.90s/trial, best loss: 751.4669168643564]2023-10-09 15:45:52,743 - build_posterior_wrapper took 0.001972 seconds\n",
      "2023-10-09 15:45:52,744 - TPE using 4/4 trials with best loss 751.466917\n",
      " 33%|███▎      | 5/15 [00:54<01:48, 10.83s/trial, best loss: 751.4669168643564]2023-10-09 15:46:03,461 - build_posterior_wrapper took 0.004995 seconds\n",
      "2023-10-09 15:46:03,462 - TPE using 5/5 trials with best loss 751.466917\n",
      " 40%|████      | 6/15 [01:08<01:47, 11.90s/trial, best loss: 751.4669168643564]2023-10-09 15:46:17,440 - build_posterior_wrapper took 0.003000 seconds\n",
      "2023-10-09 15:46:17,441 - TPE using 6/6 trials with best loss 751.466917\n",
      " 47%|████▋     | 7/15 [01:21<01:37, 12.20s/trial, best loss: 751.4669168643564]2023-10-09 15:46:30,262 - build_posterior_wrapper took 0.001997 seconds\n",
      "2023-10-09 15:46:30,263 - TPE using 7/7 trials with best loss 751.466917\n",
      " 53%|█████▎    | 8/15 [01:32<01:22, 11.73s/trial, best loss: 751.4669168643564]2023-10-09 15:46:40,962 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:46:40,962 - TPE using 8/8 trials with best loss 751.466917\n",
      " 60%|██████    | 9/15 [01:43<01:08, 11.47s/trial, best loss: 751.4669168643564]2023-10-09 15:46:51,882 - build_posterior_wrapper took 0.000975 seconds\n",
      "2023-10-09 15:46:51,883 - TPE using 9/9 trials with best loss 751.466917\n",
      " 67%|██████▋   | 10/15 [01:55<00:57, 11.59s/trial, best loss: 751.4669168643564]2023-10-09 15:47:03,727 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:47:03,728 - TPE using 10/10 trials with best loss 751.466917\n",
      " 73%|███████▎  | 11/15 [02:06<00:45, 11.38s/trial, best loss: 751.4669168643564]2023-10-09 15:47:14,628 - build_posterior_wrapper took 0.001003 seconds\n",
      "2023-10-09 15:47:14,629 - TPE using 11/11 trials with best loss 751.466917\n",
      " 80%|████████  | 12/15 [02:19<00:36, 12.08s/trial, best loss: 751.4669168643564]2023-10-09 15:47:28,322 - build_posterior_wrapper took 0.001994 seconds\n",
      "2023-10-09 15:47:28,322 - TPE using 12/12 trials with best loss 751.466917\n",
      " 87%|████████▋ | 13/15 [02:34<00:23, 11.88s/trial, best loss: 751.4669168643564]\n",
      "  0%|          | 0/15 [00:00<?, ?trial/s, best loss=?]2023-10-09 15:47:42,876 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:47:42,877 - TPE using 0 trials\n",
      "  7%|▋         | 1/15 [00:12<02:49, 12.11s/trial, best loss: 745.7470991447183]2023-10-09 15:47:54,991 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:47:54,992 - TPE using 1/1 trials with best loss 745.747099\n",
      " 13%|█▎        | 2/15 [00:24<02:37, 12.11s/trial, best loss: 745.7470991447183]2023-10-09 15:48:07,099 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:48:07,100 - TPE using 2/2 trials with best loss 745.747099\n",
      " 20%|██        | 3/15 [00:36<02:25, 12.14s/trial, best loss: 745.7470991447183]2023-10-09 15:48:19,276 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:48:19,277 - TPE using 3/3 trials with best loss 745.747099\n",
      " 27%|██▋       | 4/15 [00:48<02:13, 12.16s/trial, best loss: 745.7470991447183]2023-10-09 15:48:31,477 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:48:31,478 - TPE using 4/4 trials with best loss 745.747099\n",
      " 33%|███▎      | 5/15 [01:00<02:00, 12.10s/trial, best loss: 745.7470991447183]2023-10-09 15:48:43,454 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:48:43,455 - TPE using 5/5 trials with best loss 745.747099\n",
      " 40%|████      | 6/15 [01:11<01:45, 11.75s/trial, best loss: 745.7470991447183]2023-10-09 15:48:54,535 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:48:54,535 - TPE using 6/6 trials with best loss 745.747099\n",
      " 47%|████▋     | 7/15 [01:22<01:32, 11.60s/trial, best loss: 745.7470991447183]2023-10-09 15:49:05,823 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:49:05,823 - TPE using 7/7 trials with best loss 745.747099\n",
      " 53%|█████▎    | 8/15 [01:33<01:19, 11.41s/trial, best loss: 745.7470991447183]2023-10-09 15:49:16,825 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:49:16,826 - TPE using 8/8 trials with best loss 745.747099\n",
      " 60%|██████    | 9/15 [01:44<01:07, 11.26s/trial, best loss: 745.2440631488008]2023-10-09 15:49:27,767 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:49:27,768 - TPE using 9/9 trials with best loss 745.244063\n",
      " 67%|██████▋   | 10/15 [01:55<00:56, 11.22s/trial, best loss: 745.2440631488008]2023-10-09 15:49:38,876 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:49:38,876 - TPE using 10/10 trials with best loss 745.244063\n",
      " 73%|███████▎  | 11/15 [02:07<00:45, 11.40s/trial, best loss: 745.2440631488008]2023-10-09 15:49:50,681 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:49:50,682 - TPE using 11/11 trials with best loss 745.244063\n",
      " 80%|████████  | 12/15 [02:19<00:34, 11.40s/trial, best loss: 745.2440631488008]2023-10-09 15:50:02,104 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:50:02,105 - TPE using 12/12 trials with best loss 745.244063\n",
      " 87%|████████▋ | 13/15 [02:30<00:23, 11.56s/trial, best loss: 745.2440631488008]\n",
      "2023-10-09 15:50:13,098 - SequentialTuner - Hyperparameters optimization finished\n",
      "2023-10-09 15:50:23,987 - SequentialTuner - Return tuned graph due to the fact that obtained metric 745.244 equal or better than initial (+ 0.05% deviation) 753.418\n",
      "2023-10-09 15:50:23,988 - SequentialTuner - Final graph: {'depth': 2, 'length': 2, 'nodes': [ridge, quantile_extractor]}\n",
      "ridge - {'alpha': 9.05286146490461}\n",
      "quantile_extractor - {'window_size': 33, 'stride': 4}\n",
      "2023-10-09 15:50:23,988 - SequentialTuner - Final metric: 745.244\n",
      "Current_model - regression_pca_with_statistical_features\n",
      "2023-10-09 15:50:40,232 - DataSourceSplitter - Stratificated splitting of data is disabled.\n",
      "2023-10-09 15:50:40,233 - DataSourceSplitter - Hold out validation is applied.\n",
      "2023-10-09 15:50:40,235 - SequentialTuner - Hyperparameters optimization start: estimation of metric for initial graph\n",
      "2023-10-09 15:50:56,838 - SequentialTuner - Initial graph: {'depth': 3, 'length': 3, 'nodes': [ridge, pca, quantile_extractor]}\n",
      "ridge - {}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.9}\n",
      "quantile_extractor - {'window_size': 5} \n",
      "Initial metric: [501.166]\n",
      "  0%|          | 0/10 [00:00<?, ?trial/s, best loss=?]2023-10-09 15:50:56,846 - build_posterior_wrapper took 0.002992 seconds\n",
      "2023-10-09 15:50:56,847 - TPE using 0 trials\n",
      " 10%|█         | 1/10 [00:12<01:56, 12.98s/trial, best loss: 501.17061825960974]2023-10-09 15:51:09,820 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:51:09,821 - TPE using 1/1 trials with best loss 501.170618\n",
      " 20%|██        | 2/10 [00:27<01:50, 13.78s/trial, best loss: 501.165554177216]  2023-10-09 15:51:24,161 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:51:24,162 - TPE using 2/2 trials with best loss 501.165554\n",
      " 30%|███       | 3/10 [00:41<01:36, 13.77s/trial, best loss: 501.165554177216]2023-10-09 15:51:37,917 - build_posterior_wrapper took 0.001017 seconds\n",
      "2023-10-09 15:51:37,918 - TPE using 3/3 trials with best loss 501.165554\n",
      " 40%|████      | 4/10 [00:55<01:24, 14.12s/trial, best loss: 501.16433669497434]2023-10-09 15:51:52,576 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:51:52,576 - TPE using 4/4 trials with best loss 501.164337\n",
      " 50%|█████     | 5/10 [01:06<01:05, 13.00s/trial, best loss: 501.16433669497434]2023-10-09 15:52:03,601 - build_posterior_wrapper took 0.000969 seconds\n",
      "2023-10-09 15:52:03,602 - TPE using 5/5 trials with best loss 501.164337\n",
      " 60%|██████    | 6/10 [01:20<00:53, 13.33s/trial, best loss: 501.16433669497434]2023-10-09 15:52:17,559 - build_posterior_wrapper took 0.004003 seconds\n",
      "2023-10-09 15:52:17,559 - TPE using 6/6 trials with best loss 501.164337\n",
      " 70%|███████   | 7/10 [01:32<00:38, 12.95s/trial, best loss: 501.16433669497434]2023-10-09 15:52:29,744 - build_posterior_wrapper took 0.001021 seconds\n",
      "2023-10-09 15:52:29,745 - TPE using 7/7 trials with best loss 501.164337\n",
      " 80%|████████  | 8/10 [01:42<00:25, 12.86s/trial, best loss: 501.16433669497434]\n",
      "  0%|          | 0/10 [00:00<?, ?trial/s, best loss=?]2023-10-09 15:52:39,755 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:52:39,756 - TPE using 0 trials\n",
      " 10%|█         | 1/10 [00:13<02:02, 13.65s/trial, best loss: 501.16433669497434]2023-10-09 15:52:53,403 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:52:53,404 - TPE using 1/1 trials with best loss 501.164337\n",
      " 20%|██        | 2/10 [00:27<01:49, 13.66s/trial, best loss: 501.16433669497434]2023-10-09 15:53:07,070 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:53:07,070 - TPE using 2/2 trials with best loss 501.164337\n",
      " 30%|███       | 3/10 [00:41<01:35, 13.70s/trial, best loss: 501.16433669497434]2023-10-09 15:53:20,813 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:53:20,814 - TPE using 3/3 trials with best loss 501.164337\n",
      " 40%|████      | 4/10 [00:54<01:22, 13.72s/trial, best loss: 501.16433669497434]2023-10-09 15:53:34,572 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:53:34,573 - TPE using 4/4 trials with best loss 501.164337\n",
      " 50%|█████     | 5/10 [01:08<01:08, 13.74s/trial, best loss: 501.16433669497434]2023-10-09 15:53:48,339 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:53:48,340 - TPE using 5/5 trials with best loss 501.164337\n",
      " 60%|██████    | 6/10 [01:22<00:55, 13.96s/trial, best loss: 501.16433669497434]2023-10-09 15:54:02,739 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:54:02,740 - TPE using 6/6 trials with best loss 501.164337\n",
      " 70%|███████   | 7/10 [01:36<00:41, 13.93s/trial, best loss: 501.16433669497434]2023-10-09 15:54:16,588 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:54:16,588 - TPE using 7/7 trials with best loss 501.164337\n",
      " 80%|████████  | 8/10 [01:50<00:27, 13.84s/trial, best loss: 501.16433669497434]\n",
      "  0%|          | 0/10 [00:00<?, ?trial/s, best loss=?]2023-10-09 15:54:30,489 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:54:30,490 - TPE using 0 trials\n",
      " 10%|█         | 1/10 [00:14<02:07, 14.19s/trial, best loss: 501.1643366950236]2023-10-09 15:54:44,680 - build_posterior_wrapper took 0.001003 seconds\n",
      "2023-10-09 15:54:44,680 - TPE using 1/1 trials with best loss 501.164337\n",
      " 20%|██        | 2/10 [00:27<01:51, 13.89s/trial, best loss: 501.1643366950236]2023-10-09 15:54:58,357 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:54:58,357 - TPE using 2/2 trials with best loss 501.164337\n",
      " 30%|███       | 3/10 [00:41<01:36, 13.83s/trial, best loss: 501.1643366950236]2023-10-09 15:55:12,118 - build_posterior_wrapper took 0.001013 seconds\n",
      "2023-10-09 15:55:12,118 - TPE using 3/3 trials with best loss 501.164337\n",
      " 40%|████      | 4/10 [00:55<01:22, 13.83s/trial, best loss: 501.1643366950236]2023-10-09 15:55:25,937 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:55:25,938 - TPE using 4/4 trials with best loss 501.164337\n",
      " 50%|█████     | 5/10 [01:09<01:09, 13.85s/trial, best loss: 501.16433669499213]2023-10-09 15:55:39,820 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:55:39,821 - TPE using 5/5 trials with best loss 501.164337\n",
      " 60%|██████    | 6/10 [01:23<00:55, 13.87s/trial, best loss: 501.16433669499213]2023-10-09 15:55:53,740 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:55:53,741 - TPE using 6/6 trials with best loss 501.164337\n",
      " 70%|███████   | 7/10 [01:37<00:41, 13.97s/trial, best loss: 501.16433669499213]2023-10-09 15:56:07,907 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:56:07,908 - TPE using 7/7 trials with best loss 501.164337\n",
      " 80%|████████  | 8/10 [01:51<00:27, 13.92s/trial, best loss: 501.16433669499213]\n",
      "2023-10-09 15:56:21,858 - SequentialTuner - Hyperparameters optimization finished\n",
      "2023-10-09 15:56:36,046 - SequentialTuner - Return init graph due to the fact that obtained metric 501.164 worse than initial (+ 0.05% deviation) 500.915\n",
      "2023-10-09 15:56:36,047 - SequentialTuner - Final graph: {'depth': 3, 'length': 3, 'nodes': [ridge, pca, quantile_extractor]}\n",
      "ridge - {}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.9}\n",
      "quantile_extractor - {'window_size': 5}\n",
      "2023-10-09 15:56:36,048 - SequentialTuner - Final metric: 501.166\n",
      "Current_model - regression_with_reccurence_features\n",
      "2023-10-09 15:56:56,934 - DataSourceSplitter - Stratificated splitting of data is disabled.\n",
      "2023-10-09 15:56:56,934 - DataSourceSplitter - Hold out validation is applied.\n",
      "2023-10-09 15:56:56,936 - SequentialTuner - Hyperparameters optimization start: estimation of metric for initial graph\n",
      "2023-10-09 15:56:57,319 - SequentialTuner - Initial graph: {'depth': 2, 'length': 2, 'nodes': [ridge, recurrence_extractor]}\n",
      "ridge - {}\n",
      "recurrence_extractor - {'window_size': 20} \n",
      "Initial metric: [523.385]\n",
      "  0%|          | 0/15 [00:00<?, ?trial/s, best loss=?]2023-10-09 15:56:57,325 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:56:57,326 - TPE using 0 trials\n",
      "  7%|▋         | 1/15 [00:00<00:03,  3.56trial/s, best loss: 535.038014406195]2023-10-09 15:56:57,832 - build_posterior_wrapper took 0.226370 seconds\n",
      "2023-10-09 15:56:57,833 - TPE using 1/1 trials with best loss 535.038014\n",
      " 13%|█▎        | 2/15 [00:00<00:05,  2.59trial/s, best loss: 514.973881693618]2023-10-09 15:56:58,066 - build_posterior_wrapper took 0.001002 seconds\n",
      "2023-10-09 15:56:58,067 - TPE using 2/2 trials with best loss 514.973882\n",
      " 20%|██        | 3/15 [00:01<00:08,  1.45trial/s, best loss: 514.973881693618]2023-10-09 15:56:59,114 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:56:59,115 - TPE using 3/3 trials with best loss 514.973882\n",
      " 27%|██▋       | 4/15 [00:02<00:06,  1.71trial/s, best loss: 514.973881693618]2023-10-09 15:56:59,544 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:56:59,544 - TPE using 4/4 trials with best loss 514.973882\n",
      " 33%|███▎      | 5/15 [00:02<00:05,  1.88trial/s, best loss: 511.1691088392094]2023-10-09 15:56:59,976 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:56:59,977 - TPE using 5/5 trials with best loss 511.169109\n",
      " 40%|████      | 6/15 [00:03<00:05,  1.79trial/s, best loss: 511.1691088392094]2023-10-09 15:57:00,587 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:00,588 - TPE using 6/6 trials with best loss 511.169109\n",
      " 47%|████▋     | 7/15 [00:04<00:05,  1.52trial/s, best loss: 511.1691088392094]2023-10-09 15:57:01,454 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:01,455 - TPE using 7/7 trials with best loss 511.169109\n",
      " 53%|█████▎    | 8/15 [00:05<00:05,  1.26trial/s, best loss: 511.1691088392094]2023-10-09 15:57:02,532 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:02,532 - TPE using 8/8 trials with best loss 511.169109\n",
      " 60%|██████    | 9/15 [00:05<00:03,  1.50trial/s, best loss: 511.1691088392094]2023-10-09 15:57:02,921 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:02,921 - TPE using 9/9 trials with best loss 511.169109\n",
      " 67%|██████▋   | 10/15 [00:05<00:02,  1.87trial/s, best loss: 511.1691088392094]2023-10-09 15:57:03,166 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:03,166 - TPE using 10/10 trials with best loss 511.169109\n",
      " 73%|███████▎  | 11/15 [00:06<00:02,  1.93trial/s, best loss: 511.1691088392094]2023-10-09 15:57:03,644 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:03,645 - TPE using 11/11 trials with best loss 511.169109\n",
      " 80%|████████  | 12/15 [00:06<00:01,  1.99trial/s, best loss: 511.1691088392094]2023-10-09 15:57:04,110 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:04,110 - TPE using 12/12 trials with best loss 511.169109\n",
      " 87%|████████▋ | 13/15 [00:07<00:01,  1.61trial/s, best loss: 511.1691088392094]2023-10-09 15:57:05,003 - build_posterior_wrapper took 0.001012 seconds\n",
      "2023-10-09 15:57:05,004 - TPE using 13/13 trials with best loss 511.169109\n",
      " 93%|█████████▎| 14/15 [00:07<00:00,  1.99trial/s, best loss: 511.1691088392094]2023-10-09 15:57:05,230 - build_posterior_wrapper took 0.001000 seconds\n",
      "2023-10-09 15:57:05,231 - TPE using 14/14 trials with best loss 511.169109\n",
      "100%|██████████| 15/15 [00:08<00:00,  1.82trial/s, best loss: 511.1691088392094]\n",
      "  0%|          | 0/15 [00:00<?, ?trial/s, best loss=?]2023-10-09 15:57:05,582 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:05,583 - TPE using 0 trials\n",
      "  7%|▋         | 1/15 [00:00<00:06,  2.13trial/s, best loss: 511.02013402810684]2023-10-09 15:57:06,053 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:06,054 - TPE using 1/1 trials with best loss 511.020134\n",
      " 13%|█▎        | 2/15 [00:00<00:05,  2.21trial/s, best loss: 510.70765277559144]2023-10-09 15:57:06,495 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:06,495 - TPE using 2/2 trials with best loss 510.707653\n",
      " 20%|██        | 3/15 [00:01<00:05,  2.22trial/s, best loss: 510.30759936536884]2023-10-09 15:57:06,942 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:06,942 - TPE using 3/3 trials with best loss 510.307599\n",
      " 27%|██▋       | 4/15 [00:01<00:04,  2.23trial/s, best loss: 510.30759936536884]2023-10-09 15:57:07,388 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:07,388 - TPE using 4/4 trials with best loss 510.307599\n",
      " 33%|███▎      | 5/15 [00:02<00:04,  2.22trial/s, best loss: 510.30759936536884]2023-10-09 15:57:07,841 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:07,841 - TPE using 5/5 trials with best loss 510.307599\n",
      " 40%|████      | 6/15 [00:02<00:04,  2.22trial/s, best loss: 510.30759936536884]2023-10-09 15:57:08,291 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:08,291 - TPE using 6/6 trials with best loss 510.307599\n",
      " 47%|████▋     | 7/15 [00:03<00:03,  2.20trial/s, best loss: 510.30759936536884]2023-10-09 15:57:08,755 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:08,755 - TPE using 7/7 trials with best loss 510.307599\n",
      " 53%|█████▎    | 8/15 [00:03<00:03,  2.18trial/s, best loss: 510.30759936536884]2023-10-09 15:57:09,218 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:09,219 - TPE using 8/8 trials with best loss 510.307599\n",
      " 60%|██████    | 9/15 [00:04<00:02,  2.17trial/s, best loss: 510.30759936536884]2023-10-09 15:57:09,686 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:09,687 - TPE using 9/9 trials with best loss 510.307599\n",
      " 67%|██████▋   | 10/15 [00:04<00:02,  2.17trial/s, best loss: 510.30580299564747]2023-10-09 15:57:10,150 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:10,151 - TPE using 10/10 trials with best loss 510.305803\n",
      " 73%|███████▎  | 11/15 [00:05<00:01,  2.15trial/s, best loss: 510.30580299564747]2023-10-09 15:57:10,622 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:10,622 - TPE using 11/11 trials with best loss 510.305803\n",
      " 80%|████████  | 12/15 [00:05<00:01,  2.13trial/s, best loss: 510.30580299564747]2023-10-09 15:57:11,098 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:11,099 - TPE using 12/12 trials with best loss 510.305803\n",
      " 87%|████████▋ | 13/15 [00:05<00:00,  2.13trial/s, best loss: 510.30580299564747]2023-10-09 15:57:11,573 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:11,574 - TPE using 13/13 trials with best loss 510.305803\n",
      " 93%|█████████▎| 14/15 [00:06<00:00,  2.15trial/s, best loss: 510.30580299564747]2023-10-09 15:57:12,027 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:12,028 - TPE using 14/14 trials with best loss 510.305803\n",
      "100%|██████████| 15/15 [00:06<00:00,  2.17trial/s, best loss: 510.30580299564747]\n",
      "2023-10-09 15:57:12,488 - SequentialTuner - Hyperparameters optimization finished\n",
      "2023-10-09 15:57:12,942 - SequentialTuner - Return tuned graph due to the fact that obtained metric 510.306 equal or better than initial (+ 0.05% deviation) 523.124\n",
      "2023-10-09 15:57:12,943 - SequentialTuner - Final graph: {'depth': 2, 'length': 2, 'nodes': [ridge, recurrence_extractor]}\n",
      "ridge - {'alpha': 2.671746508377493}\n",
      "recurrence_extractor - {'window_size': 25, 'stride': 2}\n",
      "2023-10-09 15:57:12,943 - SequentialTuner - Final metric: 510.306\n",
      "Current_model - regression_pca_with_reccurence_features\n",
      "2023-10-09 15:57:13,591 - DataSourceSplitter - Stratificated splitting of data is disabled.\n",
      "2023-10-09 15:57:13,591 - DataSourceSplitter - Hold out validation is applied.\n",
      "2023-10-09 15:57:13,599 - SequentialTuner - Hyperparameters optimization start: estimation of metric for initial graph\n",
      "2023-10-09 15:57:13,983 - SequentialTuner - Initial graph: {'depth': 3, 'length': 3, 'nodes': [ridge, pca, recurrence_extractor]}\n",
      "ridge - {}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.9}\n",
      "recurrence_extractor - {'window_size': 20} \n",
      "Initial metric: [493.84]\n",
      "  0%|          | 0/10 [00:00<?, ?trial/s, best loss=?]2023-10-09 15:57:13,988 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:13,989 - TPE using 0 trials\n",
      " 10%|█         | 1/10 [00:01<00:09,  1.02s/trial, best loss: 494.48697546603415]2023-10-09 15:57:15,007 - build_posterior_wrapper took 0.001994 seconds\n",
      "2023-10-09 15:57:15,007 - TPE using 1/1 trials with best loss 494.486975\n",
      " 20%|██        | 2/10 [00:01<00:04,  1.70trial/s, best loss: 494.48697546603415]2023-10-09 15:57:15,295 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:15,295 - TPE using 2/2 trials with best loss 494.486975\n",
      " 30%|███       | 3/10 [00:01<00:03,  2.05trial/s, best loss: 494.48697546603415]2023-10-09 15:57:15,661 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:15,662 - TPE using 3/3 trials with best loss 494.486975\n",
      " 40%|████      | 4/10 [00:02<00:03,  1.58trial/s, best loss: 494.48697546603415]2023-10-09 15:57:16,520 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:16,520 - TPE using 4/4 trials with best loss 494.486975\n",
      " 50%|█████     | 5/10 [00:02<00:02,  2.03trial/s, best loss: 494.48697546603415]2023-10-09 15:57:16,764 - build_posterior_wrapper took 0.000999 seconds\n",
      "2023-10-09 15:57:16,765 - TPE using 5/5 trials with best loss 494.486975\n",
      " 60%|██████    | 6/10 [00:03<00:01,  2.08trial/s, best loss: 494.48697546603415]2023-10-09 15:57:17,220 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:17,221 - TPE using 6/6 trials with best loss 494.486975\n",
      " 70%|███████   | 7/10 [00:03<00:01,  2.12trial/s, best loss: 494.48697546603415]2023-10-09 15:57:17,674 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:17,675 - TPE using 7/7 trials with best loss 494.486975\n",
      " 80%|████████  | 8/10 [00:03<00:00,  2.58trial/s, best loss: 494.48697546603415]2023-10-09 15:57:17,879 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:17,880 - TPE using 8/8 trials with best loss 494.486975\n",
      " 90%|█████████ | 9/10 [00:04<00:00,  2.26trial/s, best loss: 494.48697546603415]2023-10-09 15:57:18,447 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:18,447 - TPE using 9/9 trials with best loss 494.486975\n",
      "100%|██████████| 10/10 [00:05<00:00,  1.97trial/s, best loss: 494.48697546603415]\n",
      "  0%|          | 0/10 [00:00<?, ?trial/s, best loss=?]2023-10-09 15:57:19,053 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:19,054 - TPE using 0 trials\n",
      " 10%|█         | 1/10 [00:01<00:09,  1.01s/trial, best loss: 510.7174688677298]2023-10-09 15:57:20,064 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:20,064 - TPE using 1/1 trials with best loss 510.717469\n",
      " 20%|██        | 2/10 [00:02<00:08,  1.01s/trial, best loss: 509.69212712866795]2023-10-09 15:57:21,078 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:21,078 - TPE using 2/2 trials with best loss 509.692127\n",
      " 30%|███       | 3/10 [00:03<00:07,  1.01s/trial, best loss: 509.69212712866795]2023-10-09 15:57:22,088 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:22,089 - TPE using 3/3 trials with best loss 509.692127\n",
      " 40%|████      | 4/10 [00:04<00:06,  1.01s/trial, best loss: 509.69212712866795]2023-10-09 15:57:23,107 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:23,108 - TPE using 4/4 trials with best loss 509.692127\n",
      " 50%|█████     | 5/10 [00:05<00:05,  1.01s/trial, best loss: 495.9498671837072] 2023-10-09 15:57:24,120 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:24,121 - TPE using 5/5 trials with best loss 495.949867\n",
      " 60%|██████    | 6/10 [00:06<00:04,  1.01s/trial, best loss: 495.9498671837072]2023-10-09 15:57:25,133 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:25,134 - TPE using 6/6 trials with best loss 495.949867\n",
      " 70%|███████   | 7/10 [00:07<00:03,  1.02s/trial, best loss: 495.9498671837072]2023-10-09 15:57:26,154 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:26,155 - TPE using 7/7 trials with best loss 495.949867\n",
      " 80%|████████  | 8/10 [00:08<00:02,  1.02s/trial, best loss: 495.9498671837072]2023-10-09 15:57:27,170 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:27,170 - TPE using 8/8 trials with best loss 495.949867\n",
      " 90%|█████████ | 9/10 [00:09<00:01,  1.01s/trial, best loss: 495.9498671837072]2023-10-09 15:57:28,180 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:28,181 - TPE using 9/9 trials with best loss 495.949867\n",
      "100%|██████████| 10/10 [00:10<00:00,  1.01s/trial, best loss: 495.9498671837072]\n",
      "  0%|          | 0/10 [00:00<?, ?trial/s, best loss=?]2023-10-09 15:57:29,199 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:29,199 - TPE using 0 trials\n",
      " 10%|█         | 1/10 [00:01<00:09,  1.02s/trial, best loss: 495.94996352423817]2023-10-09 15:57:30,218 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:30,218 - TPE using 1/1 trials with best loss 495.949964\n",
      " 20%|██        | 2/10 [00:02<00:08,  1.02s/trial, best loss: 495.94986544862127]2023-10-09 15:57:31,230 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:31,230 - TPE using 2/2 trials with best loss 495.949865\n",
      " 30%|███       | 3/10 [00:03<00:07,  1.01s/trial, best loss: 495.94986544862127]2023-10-09 15:57:32,245 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:32,245 - TPE using 3/3 trials with best loss 495.949865\n",
      " 40%|████      | 4/10 [00:04<00:06,  1.01s/trial, best loss: 495.94986544862127]2023-10-09 15:57:33,257 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:33,257 - TPE using 4/4 trials with best loss 495.949865\n",
      " 50%|█████     | 5/10 [00:05<00:05,  1.01s/trial, best loss: 495.94986544862127]2023-10-09 15:57:34,270 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:34,271 - TPE using 5/5 trials with best loss 495.949865\n",
      " 60%|██████    | 6/10 [00:06<00:04,  1.01s/trial, best loss: 495.94986544862127]2023-10-09 15:57:35,282 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:35,282 - TPE using 6/6 trials with best loss 495.949865\n",
      " 70%|███████   | 7/10 [00:07<00:03,  1.01s/trial, best loss: 495.94986544862127]2023-10-09 15:57:36,297 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:36,298 - TPE using 7/7 trials with best loss 495.949865\n",
      " 80%|████████  | 8/10 [00:08<00:02,  1.07s/trial, best loss: 495.94986544862127]2023-10-09 15:57:37,476 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:57:37,477 - TPE using 8/8 trials with best loss 495.949865\n",
      " 90%|█████████ | 9/10 [00:09<00:01,  1.05s/trial, best loss: 495.94986544862127]2023-10-09 15:57:38,488 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:38,488 - TPE using 9/9 trials with best loss 495.949865\n",
      "100%|██████████| 10/10 [00:10<00:00,  1.03s/trial, best loss: 495.94986544862127]\n",
      "2023-10-09 15:57:39,498 - SequentialTuner - Hyperparameters optimization finished\n",
      "2023-10-09 15:57:40,516 - SequentialTuner - Return init graph due to the fact that obtained metric 495.950 worse than initial (+ 0.05% deviation) 493.593\n",
      "2023-10-09 15:57:40,517 - SequentialTuner - Final graph: {'depth': 3, 'length': 3, 'nodes': [ridge, pca, recurrence_extractor]}\n",
      "ridge - {}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.9}\n",
      "recurrence_extractor - {'window_size': 20}\n",
      "2023-10-09 15:57:40,518 - SequentialTuner - Final metric: 493.840\n",
      "Current_model - regression_with_topological_features\n",
      "2023-10-09 15:57:41,069 - DataSourceSplitter - Stratificated splitting of data is disabled.\n",
      "2023-10-09 15:57:41,070 - DataSourceSplitter - Hold out validation is applied.\n",
      "2023-10-09 15:57:41,071 - SequentialTuner - Hyperparameters optimization start: estimation of metric for initial graph\n",
      "2023-10-09 15:57:49,258 - SequentialTuner - Initial graph: {'depth': 3, 'length': 3, 'nodes': [ridge, pca, topological_extractor]}\n",
      "ridge - {}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.9}\n",
      "topological_extractor - {'window_size': 20} \n",
      "Initial metric: [519.57]\n",
      "  0%|          | 0/10 [00:00<?, ?trial/s, best loss=?]2023-10-09 15:57:49,262 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:57:49,262 - TPE using 0 trials\n",
      " 10%|█         | 1/10 [00:09<01:22,  9.13s/trial, best loss: 517.6871497539621]2023-10-09 15:57:58,396 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:57:58,396 - TPE using 1/1 trials with best loss 517.687150\n",
      " 20%|██        | 2/10 [00:17<01:10,  8.76s/trial, best loss: 517.6871497539621]2023-10-09 15:58:06,889 - build_posterior_wrapper took 0.001994 seconds\n",
      "2023-10-09 15:58:06,889 - TPE using 2/2 trials with best loss 517.687150\n",
      " 30%|███       | 3/10 [00:25<00:58,  8.32s/trial, best loss: 517.6871497539621]2023-10-09 15:58:14,698 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:58:14,699 - TPE using 3/3 trials with best loss 517.687150\n",
      " 40%|████      | 4/10 [00:34<00:50,  8.42s/trial, best loss: 517.6871497539621]2023-10-09 15:58:23,273 - build_posterior_wrapper took 0.000999 seconds\n",
      "2023-10-09 15:58:23,274 - TPE using 4/4 trials with best loss 517.687150\n",
      " 50%|█████     | 5/10 [00:41<00:41,  8.20s/trial, best loss: 517.6871497539621]2023-10-09 15:58:31,088 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:58:31,089 - TPE using 5/5 trials with best loss 517.687150\n",
      " 60%|██████    | 6/10 [00:50<00:33,  8.27s/trial, best loss: 517.6871497539621]2023-10-09 15:58:39,482 - build_posterior_wrapper took 0.002992 seconds\n",
      "2023-10-09 15:58:39,483 - TPE using 6/6 trials with best loss 517.687150\n",
      " 70%|███████   | 7/10 [00:59<00:25,  8.56s/trial, best loss: 517.6871497539621]2023-10-09 15:58:48,637 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:58:48,638 - TPE using 7/7 trials with best loss 517.687150\n",
      " 80%|████████  | 8/10 [01:07<00:16,  8.35s/trial, best loss: 517.6871497539621]2023-10-09 15:58:56,546 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 15:58:56,546 - TPE using 8/8 trials with best loss 517.687150\n",
      " 90%|█████████ | 9/10 [01:16<00:08,  8.74s/trial, best loss: 516.8858520917307]2023-10-09 15:59:06,138 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:59:06,138 - TPE using 9/9 trials with best loss 516.885852\n",
      "100%|██████████| 10/10 [01:24<00:00,  8.47s/trial, best loss: 516.8858520917307]\n",
      "  0%|          | 0/10 [00:00<?, ?trial/s, best loss=?]2023-10-09 15:59:13,924 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 15:59:13,924 - TPE using 0 trials\n",
      " 10%|█         | 1/10 [00:09<01:25,  9.49s/trial, best loss: 514.4182976904648]2023-10-09 15:59:23,412 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:59:23,413 - TPE using 1/1 trials with best loss 514.418298\n",
      " 20%|██        | 2/10 [00:19<01:16,  9.51s/trial, best loss: 514.1518562778513]2023-10-09 15:59:32,934 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 15:59:32,935 - TPE using 2/2 trials with best loss 514.151856\n",
      " 30%|███       | 3/10 [00:28<01:06,  9.51s/trial, best loss: 501.8816820363014]2023-10-09 15:59:42,452 - build_posterior_wrapper took 0.001071 seconds\n",
      "2023-10-09 15:59:42,452 - TPE using 3/3 trials with best loss 501.881682\n",
      " 40%|████      | 4/10 [00:38<00:57,  9.52s/trial, best loss: 501.8816820363014]2023-10-09 15:59:51,971 - build_posterior_wrapper took 0.000969 seconds\n",
      "2023-10-09 15:59:51,971 - TPE using 4/4 trials with best loss 501.881682\n",
      " 50%|█████     | 5/10 [00:47<00:47,  9.52s/trial, best loss: 501.8816820363014]2023-10-09 16:00:01,502 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:00:01,503 - TPE using 5/5 trials with best loss 501.881682\n",
      " 60%|██████    | 6/10 [00:57<00:38,  9.56s/trial, best loss: 501.8816820363014]2023-10-09 16:00:11,124 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 16:00:11,124 - TPE using 6/6 trials with best loss 501.881682\n",
      " 70%|███████   | 7/10 [01:06<00:28,  9.54s/trial, best loss: 501.8816820363014]2023-10-09 16:00:20,640 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:00:20,641 - TPE using 7/7 trials with best loss 501.881682\n",
      " 80%|████████  | 8/10 [01:16<00:19,  9.54s/trial, best loss: 501.8816820363014]2023-10-09 16:00:30,186 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:00:30,187 - TPE using 8/8 trials with best loss 501.881682\n",
      " 90%|█████████ | 9/10 [01:25<00:09,  9.54s/trial, best loss: 501.8816820363014]2023-10-09 16:00:39,719 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:00:39,720 - TPE using 9/9 trials with best loss 501.881682\n",
      "100%|██████████| 10/10 [01:35<00:00,  9.53s/trial, best loss: 501.8816820363014]\n",
      "  0%|          | 0/10 [00:00<?, ?trial/s, best loss=?]2023-10-09 16:00:49,252 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:00:49,252 - TPE using 0 trials\n",
      " 10%|█         | 1/10 [00:09<01:26,  9.63s/trial, best loss: 501.88168204328997]2023-10-09 16:00:58,884 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:00:58,885 - TPE using 1/1 trials with best loss 501.881682\n",
      " 20%|██        | 2/10 [00:19<01:16,  9.62s/trial, best loss: 501.8816820428141] 2023-10-09 16:01:08,495 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:01:08,496 - TPE using 2/2 trials with best loss 501.881682\n",
      " 30%|███       | 3/10 [00:28<01:07,  9.65s/trial, best loss: 501.88168200999394]2023-10-09 16:01:18,179 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:01:18,180 - TPE using 3/3 trials with best loss 501.881682\n",
      " 40%|████      | 4/10 [00:39<00:59,  9.87s/trial, best loss: 501.8816819259656] 2023-10-09 16:01:28,387 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 16:01:28,387 - TPE using 4/4 trials with best loss 501.881682\n",
      " 50%|█████     | 5/10 [00:48<00:48,  9.75s/trial, best loss: 501.8816819259656]2023-10-09 16:01:37,923 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:01:37,924 - TPE using 5/5 trials with best loss 501.881682\n",
      " 60%|██████    | 6/10 [00:58<00:38,  9.67s/trial, best loss: 501.8816819259656]2023-10-09 16:01:47,448 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:01:47,449 - TPE using 6/6 trials with best loss 501.881682\n",
      " 70%|███████   | 7/10 [01:07<00:28,  9.63s/trial, best loss: 501.8816819259656]2023-10-09 16:01:56,984 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:01:56,985 - TPE using 7/7 trials with best loss 501.881682\n",
      " 80%|████████  | 8/10 [01:17<00:19,  9.62s/trial, best loss: 501.8816819259656]2023-10-09 16:02:06,602 - build_posterior_wrapper took 0.000510 seconds\n",
      "2023-10-09 16:02:06,603 - TPE using 8/8 trials with best loss 501.881682\n",
      " 90%|█████████ | 9/10 [01:26<00:09,  9.60s/trial, best loss: 501.8816819259656]2023-10-09 16:02:16,157 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:02:16,157 - TPE using 9/9 trials with best loss 501.881682\n",
      "100%|██████████| 10/10 [01:36<00:00,  9.64s/trial, best loss: 501.8816819259656]\n",
      "2023-10-09 16:02:25,692 - SequentialTuner - Hyperparameters optimization finished\n",
      "2023-10-09 16:02:35,203 - SequentialTuner - Return tuned graph due to the fact that obtained metric 501.882 equal or better than initial (+ 0.05% deviation) 519.310\n",
      "2023-10-09 16:02:35,204 - SequentialTuner - Final graph: {'depth': 3, 'length': 3, 'nodes': [ridge, pca, topological_extractor]}\n",
      "ridge - {'alpha': 8.518749615011743}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.11785781331220012}\n",
      "topological_extractor - {'window_size': 45, 'stride': 4}\n",
      "2023-10-09 16:02:35,204 - SequentialTuner - Final metric: 501.882\n",
      "Current_model - regression_pca_with_topological_features\n",
      "2023-10-09 16:02:48,739 - DataSourceSplitter - Stratificated splitting of data is disabled.\n",
      "2023-10-09 16:02:48,740 - DataSourceSplitter - Hold out validation is applied.\n",
      "2023-10-09 16:02:48,742 - SequentialTuner - Hyperparameters optimization start: estimation of metric for initial graph\n",
      "2023-10-09 16:02:56,915 - SequentialTuner - Initial graph: {'depth': 3, 'length': 3, 'nodes': [ridge, pca, topological_extractor]}\n",
      "ridge - {}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.9}\n",
      "topological_extractor - {'window_size': 20} \n",
      "Initial metric: [519.57]\n",
      "  0%|          | 0/10 [00:00<?, ?trial/s, best loss=?]2023-10-09 16:02:56,920 - build_posterior_wrapper took 0.000864 seconds\n",
      "2023-10-09 16:02:56,921 - TPE using 0 trials\n",
      " 10%|█         | 1/10 [00:07<01:10,  7.83s/trial, best loss: 523.7447801065192]2023-10-09 16:03:04,750 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 16:03:04,750 - TPE using 1/1 trials with best loss 523.744780\n",
      " 20%|██        | 2/10 [00:17<01:09,  8.63s/trial, best loss: 517.6871497539621]2023-10-09 16:03:13,942 - build_posterior_wrapper took 0.001995 seconds\n",
      "2023-10-09 16:03:13,942 - TPE using 2/2 trials with best loss 517.687150\n",
      " 30%|███       | 3/10 [00:25<01:01,  8.74s/trial, best loss: 517.0866811385202]2023-10-09 16:03:22,806 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 16:03:22,807 - TPE using 3/3 trials with best loss 517.086681\n",
      " 40%|████      | 4/10 [00:33<00:50,  8.46s/trial, best loss: 517.0866811385202]2023-10-09 16:03:30,836 - build_posterior_wrapper took 0.001996 seconds\n",
      "2023-10-09 16:03:30,837 - TPE using 4/4 trials with best loss 517.086681\n",
      " 50%|█████     | 5/10 [00:43<00:43,  8.77s/trial, best loss: 517.0866811385202]2023-10-09 16:03:40,152 - build_posterior_wrapper took 0.002993 seconds\n",
      "2023-10-09 16:03:40,152 - TPE using 5/5 trials with best loss 517.086681\n",
      " 60%|██████    | 6/10 [00:51<00:34,  8.60s/trial, best loss: 517.0866811385202]2023-10-09 16:03:48,423 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 16:03:48,424 - TPE using 6/6 trials with best loss 517.086681\n",
      " 70%|███████   | 7/10 [00:59<00:25,  8.55s/trial, best loss: 517.0866811385202]2023-10-09 16:03:56,881 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 16:03:56,882 - TPE using 7/7 trials with best loss 517.086681\n",
      " 80%|████████  | 8/10 [01:07<00:16,  8.35s/trial, best loss: 517.0866811385202]2023-10-09 16:04:04,808 - build_posterior_wrapper took 0.001995 seconds\n",
      "2023-10-09 16:04:04,808 - TPE using 8/8 trials with best loss 517.086681\n",
      " 90%|█████████ | 9/10 [01:16<00:08,  8.31s/trial, best loss: 517.0866811385202]2023-10-09 16:04:13,026 - build_posterior_wrapper took 0.001995 seconds\n",
      "2023-10-09 16:04:13,026 - TPE using 9/9 trials with best loss 517.086681\n",
      "100%|██████████| 10/10 [01:25<00:00,  8.52s/trial, best loss: 517.0866811385202]\n",
      "  0%|          | 0/10 [00:00<?, ?trial/s, best loss=?]2023-10-09 16:04:22,167 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 16:04:22,167 - TPE using 0 trials\n",
      " 10%|█         | 1/10 [00:08<01:20,  8.90s/trial, best loss: 517.7149468779435]2023-10-09 16:04:31,072 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 16:04:31,072 - TPE using 1/1 trials with best loss 517.714947\n",
      " 20%|██        | 2/10 [00:17<01:11,  8.88s/trial, best loss: 517.7149468779435]2023-10-09 16:04:39,939 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 16:04:39,939 - TPE using 2/2 trials with best loss 517.714947\n",
      " 30%|███       | 3/10 [00:26<01:02,  8.88s/trial, best loss: 501.9297349248507]2023-10-09 16:04:48,817 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 16:04:48,817 - TPE using 3/3 trials with best loss 501.929735\n",
      " 40%|████      | 4/10 [00:35<00:53,  8.88s/trial, best loss: 501.9297349248507]2023-10-09 16:04:57,695 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:04:57,696 - TPE using 4/4 trials with best loss 501.929735\n",
      " 50%|█████     | 5/10 [00:44<00:44,  8.91s/trial, best loss: 501.9297349248507]2023-10-09 16:05:06,653 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:05:06,654 - TPE using 5/5 trials with best loss 501.929735\n",
      " 60%|██████    | 6/10 [00:53<00:35,  8.89s/trial, best loss: 501.9297349248507]2023-10-09 16:05:15,522 - build_posterior_wrapper took 0.000997 seconds\n",
      "2023-10-09 16:05:15,522 - TPE using 6/6 trials with best loss 501.929735\n",
      " 70%|███████   | 7/10 [01:02<00:26,  8.88s/trial, best loss: 501.9297349248507]2023-10-09 16:05:24,385 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 16:05:24,385 - TPE using 7/7 trials with best loss 501.929735\n",
      " 80%|████████  | 8/10 [01:11<00:17,  8.89s/trial, best loss: 501.9297349248507]2023-10-09 16:05:33,300 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:05:33,301 - TPE using 8/8 trials with best loss 501.929735\n",
      " 90%|█████████ | 9/10 [01:20<00:08,  8.89s/trial, best loss: 501.9297349248507]2023-10-09 16:05:42,193 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 16:05:42,193 - TPE using 9/9 trials with best loss 501.929735\n",
      "100%|██████████| 10/10 [01:28<00:00,  8.89s/trial, best loss: 501.9297349248507]\n",
      "  0%|          | 0/10 [00:00<?, ?trial/s, best loss=?]2023-10-09 16:05:51,073 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 16:05:51,073 - TPE using 0 trials\n",
      " 10%|█         | 1/10 [00:09<01:21,  9.08s/trial, best loss: 501.9297349605441]2023-10-09 16:06:00,158 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:06:00,159 - TPE using 1/1 trials with best loss 501.929735\n",
      " 20%|██        | 2/10 [00:18<01:14,  9.29s/trial, best loss: 501.92973492425574]2023-10-09 16:06:09,594 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:06:09,595 - TPE using 2/2 trials with best loss 501.929735\n",
      " 30%|███       | 3/10 [00:27<01:03,  9.10s/trial, best loss: 501.92973492425574]2023-10-09 16:06:18,460 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:06:18,461 - TPE using 3/3 trials with best loss 501.929735\n",
      " 40%|████      | 4/10 [00:36<00:54,  9.01s/trial, best loss: 501.92973492425574]2023-10-09 16:06:27,336 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:06:27,336 - TPE using 4/4 trials with best loss 501.929735\n",
      " 50%|█████     | 5/10 [00:45<00:44,  8.96s/trial, best loss: 501.92973492425574]2023-10-09 16:06:36,217 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:06:36,217 - TPE using 5/5 trials with best loss 501.929735\n",
      " 60%|██████    | 6/10 [00:54<00:35,  8.93s/trial, best loss: 501.92973492425574]2023-10-09 16:06:45,077 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:06:45,077 - TPE using 6/6 trials with best loss 501.929735\n",
      " 70%|███████   | 7/10 [01:02<00:26,  8.91s/trial, best loss: 501.92973492425574]2023-10-09 16:06:53,953 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:06:53,954 - TPE using 7/7 trials with best loss 501.929735\n",
      " 80%|████████  | 8/10 [01:11<00:17,  8.90s/trial, best loss: 501.92973492425574]2023-10-09 16:07:02,820 - build_posterior_wrapper took 0.000000 seconds\n",
      "2023-10-09 16:07:02,821 - TPE using 8/8 trials with best loss 501.929735\n",
      " 90%|█████████ | 9/10 [01:20<00:08,  8.92s/trial, best loss: 501.92973492425574]2023-10-09 16:07:11,780 - build_posterior_wrapper took 0.000998 seconds\n",
      "2023-10-09 16:07:11,780 - TPE using 9/9 trials with best loss 501.929735\n",
      "100%|██████████| 10/10 [01:29<00:00,  8.96s/trial, best loss: 501.92973492425574]\n",
      "2023-10-09 16:07:20,644 - SequentialTuner - Hyperparameters optimization finished\n",
      "2023-10-09 16:07:29,528 - SequentialTuner - Return tuned graph due to the fact that obtained metric 501.930 equal or better than initial (+ 0.05% deviation) 519.310\n",
      "2023-10-09 16:07:29,529 - SequentialTuner - Final graph: {'depth': 3, 'length': 3, 'nodes': [ridge, pca, topological_extractor]}\n",
      "ridge - {'alpha': 0.9147556926771431}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.27530461315055466}\n",
      "topological_extractor - {'window_size': 35, 'stride': 1}\n",
      "2023-10-09 16:07:29,530 - SequentialTuner - Final metric: 501.930\n"
     ]
    }
   ],
   "source": [
    "with IndustrialModels():\n",
    "    for model in model_dict.keys():\n",
    "        print(f'Current_model - {model}')\n",
    "        pipeline = model_dict[model].build()\n",
    "        tuned_pipeline = tuning_industrial_pipelines(pipeline,tuning_params,train_data)\n",
    "        tuned_pipeline.fit(input_data)\n",
    "        features = tuned_pipeline.predict(val_data).predict\n",
    "        metric = calculate_regression_metric(test_target=test_data[1], labels=features)\n",
    "        metric = metric.T\n",
    "        metric.columns = metric.columns.values\n",
    "        metric['model_params'] = metric['model_params'] = str({node:node.parameters for node in tuned_pipeline.graph_description['nodes']})\n",
    "        metric_dict.update({model: metric})"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "outputs": [
    {
     "data": {
      "text/plain": "                                         r2_score: mean_squared_error:  \\\nregression_with_reccurence_features       0.085962       241398.503003   \nregression_pca_with_reccurence_features   0.080487       242844.478589   \nregression_with_topological_features      0.060005       248253.725979   \nregression_pca_with_topological_features  0.059872       248288.864243   \nregression_pca_with_statistical_features -0.002798       264840.107063   \nregression_with_statistical_features     -1.483146       655801.748375   \n\n                                         root_mean_squared_error:  \\\nregression_with_reccurence_features                    491.323216   \nregression_pca_with_reccurence_features                492.792531   \nregression_with_topological_features                   498.250666   \nregression_pca_with_topological_features               498.285926   \nregression_pca_with_statistical_features               514.626182   \nregression_with_statistical_features                   809.815873   \n\n                                         mean_absolute_error  \\\nregression_with_reccurence_features               269.058427   \nregression_pca_with_reccurence_features           244.728677   \nregression_with_topological_features              240.835807   \nregression_pca_with_topological_features          244.145529   \nregression_pca_with_statistical_features          261.164925   \nregression_with_statistical_features              560.532275   \n\n                                         median_absolute_error  \\\nregression_with_reccurence_features                 144.576028   \nregression_pca_with_reccurence_features             120.852159   \nregression_with_topological_features                116.727277   \nregression_pca_with_topological_features            121.311948   \nregression_pca_with_statistical_features            133.701291   \nregression_with_statistical_features                360.750976   \n\n                                         explained_variance_score  \\\nregression_with_reccurence_features                      0.087197   \nregression_pca_with_reccurence_features                  0.085936   \nregression_with_topological_features                     0.062829   \nregression_pca_with_topological_features                 0.061968   \nregression_pca_with_statistical_features                -0.001518   \nregression_with_statistical_features                    -1.434199   \n\n                                            max_error d2_absolute_error_score  \\\nregression_with_reccurence_features       2108.264973               -0.394872   \nregression_pca_with_reccurence_features   2167.000715                -0.26874   \nregression_with_topological_features      2076.191042               -0.248559   \nregression_pca_with_topological_features  2139.069571               -0.265717   \nregression_pca_with_statistical_features  2337.330582                -0.35395   \nregression_with_statistical_features      2405.811682               -1.905952   \n\n                                                                               model_params  \nregression_with_reccurence_features       {ridge: {'alpha': 2.671746508377493}, recurren...  \nregression_pca_with_reccurence_features   {ridge: {}, pca: {'svd_solver': 'full', 'n_com...  \nregression_with_topological_features      {ridge: {'alpha': 8.518749615011743}, pca: {'s...  \nregression_pca_with_topological_features  {ridge: {'alpha': 0.9147556926771431}, pca: {'...  \nregression_pca_with_statistical_features  {ridge: {}, pca: {'svd_solver': 'full', 'n_com...  \nregression_with_statistical_features      {ridge: {'alpha': 9.05286146490461}, quantile_...  ",
      "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>r2_score:</th>\n      <th>mean_squared_error:</th>\n      <th>root_mean_squared_error:</th>\n      <th>mean_absolute_error</th>\n      <th>median_absolute_error</th>\n      <th>explained_variance_score</th>\n      <th>max_error</th>\n      <th>d2_absolute_error_score</th>\n      <th>model_params</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>regression_with_reccurence_features</th>\n      <td>0.085962</td>\n      <td>241398.503003</td>\n      <td>491.323216</td>\n      <td>269.058427</td>\n      <td>144.576028</td>\n      <td>0.087197</td>\n      <td>2108.264973</td>\n      <td>-0.394872</td>\n      <td>{ridge: {'alpha': 2.671746508377493}, recurren...</td>\n    </tr>\n    <tr>\n      <th>regression_pca_with_reccurence_features</th>\n      <td>0.080487</td>\n      <td>242844.478589</td>\n      <td>492.792531</td>\n      <td>244.728677</td>\n      <td>120.852159</td>\n      <td>0.085936</td>\n      <td>2167.000715</td>\n      <td>-0.26874</td>\n      <td>{ridge: {}, pca: {'svd_solver': 'full', 'n_com...</td>\n    </tr>\n    <tr>\n      <th>regression_with_topological_features</th>\n      <td>0.060005</td>\n      <td>248253.725979</td>\n      <td>498.250666</td>\n      <td>240.835807</td>\n      <td>116.727277</td>\n      <td>0.062829</td>\n      <td>2076.191042</td>\n      <td>-0.248559</td>\n      <td>{ridge: {'alpha': 8.518749615011743}, pca: {'s...</td>\n    </tr>\n    <tr>\n      <th>regression_pca_with_topological_features</th>\n      <td>0.059872</td>\n      <td>248288.864243</td>\n      <td>498.285926</td>\n      <td>244.145529</td>\n      <td>121.311948</td>\n      <td>0.061968</td>\n      <td>2139.069571</td>\n      <td>-0.265717</td>\n      <td>{ridge: {'alpha': 0.9147556926771431}, pca: {'...</td>\n    </tr>\n    <tr>\n      <th>regression_pca_with_statistical_features</th>\n      <td>-0.002798</td>\n      <td>264840.107063</td>\n      <td>514.626182</td>\n      <td>261.164925</td>\n      <td>133.701291</td>\n      <td>-0.001518</td>\n      <td>2337.330582</td>\n      <td>-0.35395</td>\n      <td>{ridge: {}, pca: {'svd_solver': 'full', 'n_com...</td>\n    </tr>\n    <tr>\n      <th>regression_with_statistical_features</th>\n      <td>-1.483146</td>\n      <td>655801.748375</td>\n      <td>809.815873</td>\n      <td>560.532275</td>\n      <td>360.750976</td>\n      <td>-1.434199</td>\n      <td>2405.811682</td>\n      <td>-1.905952</td>\n      <td>{ridge: {'alpha': 9.05286146490461}, quantile_...</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_tuned = pd.concat([x for x in metric_dict.values()],axis=0).T\n",
    "df_tuned.columns = list(metric_dict.keys())\n",
    "df_tuned.T.sort_values('root_mean_squared_error:')"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "outputs": [
    {
     "data": {
      "text/plain": "\"{ridge: {'alpha': 2.671746508377493}, recurrence_extractor: {'window_size': 25, 'stride': 2}}\""
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_tuned.T.sort_values('root_mean_squared_error:')['model_params'].iloc[0]"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "outputs": [],
   "source": [
    "best_tuned = df_tuned.T.sort_values('root_mean_squared_error:')['root_mean_squared_error:'].iloc[0]"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Even better? AutoML approach"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 16:24:50,258 - AssumptionsHandler - Memory consumption for fitting of the initial pipeline in main session: current 0.8 MiB, max: 5.6 MiB\n",
      "2023-10-09 16:24:50,259 - ApiComposer - Initial pipeline was fitted in 1.9 sec.\n",
      "2023-10-09 16:24:50,260 - AssumptionsHandler - Preset was changed to best_quality due to fit time estimation for initial model.\n",
      "2023-10-09 16:24:50,267 - ApiComposer - AutoML configured. Parameters tuning: False. Time limit: 10 min. Set of candidate models: ['rfr', 'ridge', 'scaling', 'normalization', 'pca', 'catboostreg', 'xgbreg', 'svr', 'dtreg', 'treg', 'knnreg', 'fast_ica', 'kernel_pca', 'isolation_forest_reg', 'rfe_lin_reg', 'rfe_non_lin_reg'].\n",
      "2023-10-09 16:24:50,270 - ApiComposer - Pipeline composition started.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generations:   0%|          | 1/10000 [00:00<?, ?gen/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 16:24:57,550 - MultiprocessingDispatcher - 2 individuals out of 2 in previous population were evaluated successfully.\n",
      "2023-10-09 16:25:19,284 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 16:25:27,212 - MultiprocessingDispatcher - 12 individuals out of 12 in previous population were evaluated successfully.\n",
      "2023-10-09 16:25:39,878 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 16:25:41,586 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:25:48,811 - MultiprocessingDispatcher - 13 individuals out of 13 in previous population were evaluated successfully.\n",
      "2023-10-09 16:25:50,694 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:25:51,165 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:25:58,247 - MultiprocessingDispatcher - 20 individuals out of 20 in previous population were evaluated successfully.\n",
      "2023-10-09 16:26:04,076 - MultiprocessingDispatcher - 14 individuals out of 14 in previous population were evaluated successfully.\n",
      "2023-10-09 16:26:23,525 - MultiprocessingDispatcher - 25 individuals out of 25 in previous population were evaluated successfully.\n",
      "2023-10-09 16:26:35,125 - MultiprocessingDispatcher - 29 individuals out of 29 in previous population were evaluated successfully.\n",
      "2023-10-09 16:26:35,147 - MultiprocessingDispatcher - 33 individuals out of 33 in previous population were evaluated successfully.\n",
      "2023-10-09 16:26:55,171 - MultiprocessingDispatcher - 26 individuals out of 26 in previous population were evaluated successfully.\n",
      "2023-10-09 16:26:57,690 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 16:27:15,193 - MultiprocessingDispatcher - 27 individuals out of 27 in previous population were evaluated successfully.\n",
      "2023-10-09 16:27:29,972 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 16:27:45,481 - MultiprocessingDispatcher - 27 individuals out of 27 in previous population were evaluated successfully.\n",
      "2023-10-09 16:28:04,434 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 16:28:19,779 - MultiprocessingDispatcher - 28 individuals out of 28 in previous population were evaluated successfully.\n",
      "2023-10-09 16:28:22,130 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 16:28:34,909 - MultiprocessingDispatcher - 31 individuals out of 31 in previous population were evaluated successfully.\n",
      "2023-10-09 16:28:53,840 - MultiprocessingDispatcher - 22 individuals out of 22 in previous population were evaluated successfully.\n",
      "2023-10-09 16:28:53,868 - MultiprocessingDispatcher - 47 individuals out of 47 in previous population were evaluated successfully.\n",
      "2023-10-09 16:29:10,437 - MultiprocessingDispatcher - 43 individuals out of 43 in previous population were evaluated successfully.\n",
      "2023-10-09 16:29:37,839 - MultiprocessingDispatcher - 40 individuals out of 41 in previous population were evaluated successfully.\n",
      "2023-10-09 16:29:39,929 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 16:29:54,690 - MultiprocessingDispatcher - 39 individuals out of 39 in previous population were evaluated successfully.\n",
      "2023-10-09 16:30:01,097 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 16:30:12,245 - MultiprocessingDispatcher - 36 individuals out of 36 in previous population were evaluated successfully.\n",
      "2023-10-09 16:30:17,044 - MultiprocessingDispatcher - 12 individuals out of 12 in previous population were evaluated successfully.\n",
      "2023-10-09 16:30:51,122 - MultiprocessingDispatcher - 37 individuals out of 37 in previous population were evaluated successfully.\n",
      "2023-10-09 16:31:00,067 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 16:31:00,096 - MultiprocessingDispatcher - 41 individuals out of 41 in previous population were evaluated successfully.\n",
      "2023-10-09 16:31:15,254 - MultiprocessingDispatcher - 30 individuals out of 30 in previous population were evaluated successfully.\n",
      "2023-10-09 16:31:22,064 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 16:31:42,785 - MultiprocessingDispatcher - 31 individuals out of 31 in previous population were evaluated successfully.\n",
      "2023-10-09 16:31:45,923 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 16:32:08,025 - MultiprocessingDispatcher - 37 individuals out of 37 in previous population were evaluated successfully.\n",
      "2023-10-09 16:32:23,800 - MultiprocessingDispatcher - 34 individuals out of 34 in previous population were evaluated successfully.\n",
      "2023-10-09 16:32:30,219 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 16:32:42,784 - MultiprocessingDispatcher - 31 individuals out of 31 in previous population were evaluated successfully.\n",
      "2023-10-09 16:32:46,942 - MultiprocessingDispatcher - 11 individuals out of 11 in previous population were evaluated successfully.\n",
      "2023-10-09 16:32:46,969 - MultiprocessingDispatcher - 38 individuals out of 38 in previous population were evaluated successfully.\n",
      "2023-10-09 16:33:12,788 - MultiprocessingDispatcher - 30 individuals out of 30 in previous population were evaluated successfully.\n",
      "2023-10-09 16:33:16,688 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 16:33:29,013 - MultiprocessingDispatcher - 27 individuals out of 27 in previous population were evaluated successfully.\n",
      "2023-10-09 16:33:35,993 - MultiprocessingDispatcher - 9 individuals out of 9 in previous population were evaluated successfully.\n",
      "2023-10-09 16:33:55,014 - MultiprocessingDispatcher - 25 individuals out of 25 in previous population were evaluated successfully.\n",
      "2023-10-09 16:34:03,141 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 16:34:32,023 - MultiprocessingDispatcher - 24 individuals out of 24 in previous population were evaluated successfully.\n",
      "2023-10-09 16:34:42,811 - MultiprocessingDispatcher - 24 individuals out of 24 in previous population were evaluated successfully.\n",
      "2023-10-09 16:34:46,103 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 16:34:46,330 - GroupedCondition - Optimisation stopped: Time limit is reached\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generations:   0%|          | 1/10000 [09:56<?, ?gen/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 16:34:46,501 - ApiComposer - Model generation finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 16:34:49,065 - FEDOT logger - Final pipeline was fitted\n",
      "2023-10-09 16:34:49,066 - FEDOT logger - Final pipeline: {'depth': 7, 'length': 7, 'nodes': [treg, svr, scaling, isolation_forest_reg, resample, fast_ica, pca]}\n",
      "treg - {'max_features': 0.9692738372005898, 'min_samples_split': 18, 'min_samples_leaf': 11, 'bootstrap': False}\n",
      "svr - {}\n",
      "scaling - {}\n",
      "isolation_forest_reg - {}\n",
      "resample - {'balance': 'expand_minority', 'replace': False, 'balance_ratio': 1}\n",
      "fast_ica - {'whiten': 'unit-variance'}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.7}\n",
      "2023-10-09 16:34:49,068 - MemoryAnalytics - Memory consumption for finish in main session: current 4.7 MiB, max: 10.8 MiB\n",
      "2023-10-09 16:34:53,903 - AssumptionsHandler - Memory consumption for fitting of the initial pipeline in main session: current 0.7 MiB, max: 5.4 MiB\n",
      "2023-10-09 16:34:53,904 - ApiComposer - Initial pipeline was fitted in 1.6 sec.\n",
      "2023-10-09 16:34:53,905 - AssumptionsHandler - Preset was changed to best_quality due to fit time estimation for initial model.\n",
      "2023-10-09 16:34:53,911 - ApiComposer - AutoML configured. Parameters tuning: False. Time limit: 10 min. Set of candidate models: ['rfr', 'ridge', 'scaling', 'normalization', 'pca', 'catboostreg', 'xgbreg', 'svr', 'dtreg', 'treg', 'knnreg', 'fast_ica', 'kernel_pca', 'isolation_forest_reg', 'rfe_lin_reg', 'rfe_non_lin_reg'].\n",
      "2023-10-09 16:34:53,913 - ApiComposer - Pipeline composition started.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generations:   0%|          | 1/10000 [00:00<?, ?gen/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 16:34:56,120 - MultiprocessingDispatcher - 2 individuals out of 2 in previous population were evaluated successfully.\n",
      "2023-10-09 16:35:21,001 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 16:35:50,180 - MultiprocessingDispatcher - 17 individuals out of 17 in previous population were evaluated successfully.\n",
      "2023-10-09 16:35:50,915 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:35:53,723 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 16:35:54,380 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:00,045 - MultiprocessingDispatcher - 12 individuals out of 12 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:03,509 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:03,956 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:05,822 - MultiprocessingDispatcher - 14 individuals out of 14 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:11,090 - MultiprocessingDispatcher - 12 individuals out of 12 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:17,325 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:17,961 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:17,975 - MultiprocessingDispatcher - 19 individuals out of 19 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:20,367 - MultiprocessingDispatcher - 13 individuals out of 13 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:23,960 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:24,359 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:30,897 - MultiprocessingDispatcher - 15 individuals out of 15 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:31,502 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:33,328 - MultiprocessingDispatcher - 12 individuals out of 12 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:35,121 - MultiprocessingDispatcher - 14 individuals out of 14 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:37,298 - MultiprocessingDispatcher - 14 individuals out of 14 in previous population were evaluated successfully.\n",
      "2023-10-09 16:36:39,834 - MultiprocessingDispatcher - 11 individuals out of 11 in previous population were evaluated successfully.\n",
      "2023-10-09 16:37:04,628 - MultiprocessingDispatcher - 35 individuals out of 35 in previous population were evaluated successfully.\n",
      "2023-10-09 16:37:16,053 - MultiprocessingDispatcher - 22 individuals out of 22 in previous population were evaluated successfully.\n",
      "2023-10-09 16:37:39,834 - MultiprocessingDispatcher - 46 individuals out of 46 in previous population were evaluated successfully.\n",
      "2023-10-09 16:37:41,477 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 16:37:41,500 - MultiprocessingDispatcher - 48 individuals out of 48 in previous population were evaluated successfully.\n",
      "2023-10-09 16:37:59,986 - MultiprocessingDispatcher - 43 individuals out of 43 in previous population were evaluated successfully.\n",
      "2023-10-09 16:38:04,098 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 16:38:11,357 - MultiprocessingDispatcher - 43 individuals out of 43 in previous population were evaluated successfully.\n",
      "2023-10-09 16:38:12,382 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 16:38:20,270 - MultiprocessingDispatcher - 39 individuals out of 39 in previous population were evaluated successfully.\n",
      "2023-10-09 16:38:21,781 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 16:38:29,797 - MultiprocessingDispatcher - 39 individuals out of 40 in previous population were evaluated successfully.\n",
      "2023-10-09 16:38:31,404 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 16:38:40,511 - MultiprocessingDispatcher - 45 individuals out of 45 in previous population were evaluated successfully.\n",
      "2023-10-09 16:38:40,532 - MultiprocessingDispatcher - 31 individuals out of 31 in previous population were evaluated successfully.\n",
      "2023-10-09 16:38:56,503 - MultiprocessingDispatcher - 28 individuals out of 28 in previous population were evaluated successfully.\n",
      "2023-10-09 16:39:05,242 - MultiprocessingDispatcher - 29 individuals out of 29 in previous population were evaluated successfully.\n",
      "2023-10-09 16:39:23,815 - MultiprocessingDispatcher - 26 individuals out of 27 in previous population were evaluated successfully.\n",
      "2023-10-09 16:39:25,149 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:39:44,176 - MultiprocessingDispatcher - 25 individuals out of 25 in previous population were evaluated successfully.\n",
      "2023-10-09 16:39:46,703 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 16:39:56,989 - MultiprocessingDispatcher - 27 individuals out of 28 in previous population were evaluated successfully.\n",
      "2023-10-09 16:39:58,424 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:39:58,438 - MultiprocessingDispatcher - 28 individuals out of 28 in previous population were evaluated successfully.\n",
      "2023-10-09 16:40:15,763 - MultiprocessingDispatcher - 25 individuals out of 27 in previous population were evaluated successfully.\n",
      "2023-10-09 16:40:20,110 - MultiprocessingDispatcher - 2 individuals out of 2 in previous population were evaluated successfully.\n",
      "2023-10-09 16:40:34,822 - MultiprocessingDispatcher - 26 individuals out of 26 in previous population were evaluated successfully.\n",
      "2023-10-09 16:40:44,415 - MultiprocessingDispatcher - 24 individuals out of 27 in previous population were evaluated successfully.\n",
      "2023-10-09 16:40:45,834 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:41:04,394 - MultiprocessingDispatcher - 26 individuals out of 26 in previous population were evaluated successfully.\n",
      "2023-10-09 16:41:15,072 - MultiprocessingDispatcher - 22 individuals out of 26 in previous population were evaluated successfully.\n",
      "2023-10-09 16:41:16,319 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 16:41:16,333 - MultiprocessingDispatcher - 26 individuals out of 26 in previous population were evaluated successfully.\n",
      "2023-10-09 16:41:28,292 - MultiprocessingDispatcher - 21 individuals out of 23 in previous population were evaluated successfully.\n",
      "2023-10-09 16:41:29,914 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:41:37,418 - MultiprocessingDispatcher - 24 individuals out of 24 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:02,028 - MultiprocessingDispatcher - 25 individuals out of 25 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:06,474 - MultiprocessingDispatcher - 20 individuals out of 20 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:07,951 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:14,565 - MultiprocessingDispatcher - 25 individuals out of 25 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:14,577 - MultiprocessingDispatcher - 23 individuals out of 23 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:22,104 - MultiprocessingDispatcher - 18 individuals out of 18 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:22,925 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:27,272 - MultiprocessingDispatcher - 22 individuals out of 22 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:32,081 - MultiprocessingDispatcher - 19 individuals out of 19 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:33,430 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:38,539 - MultiprocessingDispatcher - 20 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:39,745 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:56,537 - MultiprocessingDispatcher - 20 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:57,735 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:42:57,747 - MultiprocessingDispatcher - 20 individuals out of 20 in previous population were evaluated successfully.\n",
      "2023-10-09 16:43:05,387 - MultiprocessingDispatcher - 16 individuals out of 16 in previous population were evaluated successfully.\n",
      "2023-10-09 16:43:07,125 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:43:12,284 - MultiprocessingDispatcher - 19 individuals out of 19 in previous population were evaluated successfully.\n",
      "2023-10-09 16:43:20,740 - MultiprocessingDispatcher - 18 individuals out of 19 in previous population were evaluated successfully.\n",
      "2023-10-09 16:43:27,405 - MultiprocessingDispatcher - 19 individuals out of 19 in previous population were evaluated successfully.\n",
      "2023-10-09 16:43:44,861 - MultiprocessingDispatcher - 20 individuals out of 20 in previous population were evaluated successfully.\n",
      "2023-10-09 16:43:44,872 - MultiprocessingDispatcher - 19 individuals out of 19 in previous population were evaluated successfully.\n",
      "2023-10-09 16:44:13,537 - MultiprocessingDispatcher - 17 individuals out of 17 in previous population were evaluated successfully.\n",
      "2023-10-09 16:44:15,009 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:44:34,381 - MultiprocessingDispatcher - 17 individuals out of 17 in previous population were evaluated successfully.\n",
      "2023-10-09 16:44:35,966 - MultiprocessingDispatcher - 3 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:44:40,858 - MultiprocessingDispatcher - 18 individuals out of 18 in previous population were evaluated successfully.\n",
      "2023-10-09 16:44:40,927 - GroupedCondition - Optimisation stopped: Time limit is reached\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generations:   0%|          | 1/10000 [09:47<?, ?gen/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 16:44:41,148 - ApiComposer - Model generation finished\n",
      "2023-10-09 16:44:42,161 - FEDOT logger - Final pipeline was fitted\n",
      "2023-10-09 16:44:42,163 - FEDOT logger - Final pipeline: {'depth': 5, 'length': 5, 'nodes': [ridge, pca, isolation_forest_reg, normalization, resample]}\n",
      "ridge - {'alpha': 6.62055874397956}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.2187848583572159}\n",
      "isolation_forest_reg - {}\n",
      "normalization - {}\n",
      "resample - {'balance': 'expand_minority', 'replace': False, 'balance_ratio': 1}\n",
      "2023-10-09 16:44:42,163 - MemoryAnalytics - Memory consumption for finish in main session: current 4.1 MiB, max: 10.0 MiB\n",
      "2023-10-09 16:44:46,773 - AssumptionsHandler - Memory consumption for fitting of the initial pipeline in main session: current 0.7 MiB, max: 5.4 MiB\n",
      "2023-10-09 16:44:46,774 - ApiComposer - Initial pipeline was fitted in 1.6 sec.\n",
      "2023-10-09 16:44:46,774 - AssumptionsHandler - Preset was changed to best_quality due to fit time estimation for initial model.\n",
      "2023-10-09 16:44:46,780 - ApiComposer - AutoML configured. Parameters tuning: False. Time limit: 10 min. Set of candidate models: ['rfr', 'ridge', 'scaling', 'normalization', 'pca', 'catboostreg', 'xgbreg', 'svr', 'dtreg', 'treg', 'knnreg', 'fast_ica', 'kernel_pca', 'isolation_forest_reg', 'rfe_lin_reg', 'rfe_non_lin_reg'].\n",
      "2023-10-09 16:44:46,782 - ApiComposer - Pipeline composition started.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generations:   0%|          | 1/10000 [00:00<?, ?gen/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 16:44:48,477 - MultiprocessingDispatcher - 2 individuals out of 2 in previous population were evaluated successfully.\n",
      "2023-10-09 16:45:21,106 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 16:45:56,287 - MultiprocessingDispatcher - 15 individuals out of 15 in previous population were evaluated successfully.\n",
      "2023-10-09 16:46:01,450 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 16:46:42,335 - MultiprocessingDispatcher - 17 individuals out of 17 in previous population were evaluated successfully.\n",
      "2023-10-09 16:46:44,233 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:46:51,971 - MultiprocessingDispatcher - 20 individuals out of 20 in previous population were evaluated successfully.\n",
      "2023-10-09 16:46:56,211 - MultiprocessingDispatcher - 14 individuals out of 14 in previous population were evaluated successfully.\n",
      "2023-10-09 16:47:09,117 - MultiprocessingDispatcher - 25 individuals out of 25 in previous population were evaluated successfully.\n",
      "2023-10-09 16:47:20,564 - MultiprocessingDispatcher - 24 individuals out of 24 in previous population were evaluated successfully.\n",
      "2023-10-09 16:47:23,720 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:47:23,742 - MultiprocessingDispatcher - 31 individuals out of 31 in previous population were evaluated successfully.\n",
      "2023-10-09 16:47:35,857 - MultiprocessingDispatcher - 23 individuals out of 24 in previous population were evaluated successfully.\n",
      "2023-10-09 16:47:40,745 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 16:47:59,023 - MultiprocessingDispatcher - 22 individuals out of 24 in previous population were evaluated successfully.\n",
      "2023-10-09 16:48:04,655 - MultiprocessingDispatcher - 9 individuals out of 9 in previous population were evaluated successfully.\n",
      "2023-10-09 16:48:16,263 - MultiprocessingDispatcher - 28 individuals out of 29 in previous population were evaluated successfully.\n",
      "2023-10-09 16:48:28,098 - MultiprocessingDispatcher - 27 individuals out of 27 in previous population were evaluated successfully.\n",
      "2023-10-09 16:48:31,933 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:48:50,662 - MultiprocessingDispatcher - 28 individuals out of 28 in previous population were evaluated successfully.\n",
      "2023-10-09 16:48:50,675 - MultiprocessingDispatcher - 29 individuals out of 29 in previous population were evaluated successfully.\n",
      "2023-10-09 16:49:07,139 - MultiprocessingDispatcher - 24 individuals out of 24 in previous population were evaluated successfully.\n",
      "2023-10-09 16:49:24,134 - MultiprocessingDispatcher - 3 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:50:01,994 - MultiprocessingDispatcher - 27 individuals out of 28 in previous population were evaluated successfully.\n",
      "2023-10-09 16:50:12,834 - MultiprocessingDispatcher - 25 individuals out of 25 in previous population were evaluated successfully.\n",
      "2023-10-09 16:50:14,884 - MultiprocessingDispatcher - 2 individuals out of 2 in previous population were evaluated successfully.\n",
      "2023-10-09 16:50:33,650 - MultiprocessingDispatcher - 24 individuals out of 24 in previous population were evaluated successfully.\n",
      "2023-10-09 16:50:34,546 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:50:53,164 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 16:50:55,176 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 16:50:55,191 - MultiprocessingDispatcher - 25 individuals out of 25 in previous population were evaluated successfully.\n",
      "2023-10-09 16:51:03,868 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 16:51:04,813 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:51:09,681 - MultiprocessingDispatcher - 19 individuals out of 19 in previous population were evaluated successfully.\n",
      "2023-10-09 16:51:17,501 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 16:51:30,961 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 16:51:31,926 - MultiprocessingDispatcher - 2 individuals out of 2 in previous population were evaluated successfully.\n",
      "2023-10-09 16:51:37,404 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 16:51:38,910 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:51:44,517 - MultiprocessingDispatcher - 16 individuals out of 16 in previous population were evaluated successfully.\n",
      "2023-10-09 16:51:51,744 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 16:51:51,756 - MultiprocessingDispatcher - 24 individuals out of 24 in previous population were evaluated successfully.\n",
      "2023-10-09 16:52:10,245 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 16:52:11,816 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:52:21,456 - MultiprocessingDispatcher - 17 individuals out of 17 in previous population were evaluated successfully.\n",
      "2023-10-09 16:52:22,793 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 16:52:39,366 - MultiprocessingDispatcher - 16 individuals out of 16 in previous population were evaluated successfully.\n",
      "2023-10-09 16:52:41,891 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 16:52:59,225 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:00,212 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:09,078 - MultiprocessingDispatcher - 17 individuals out of 18 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:10,410 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:12,124 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:12,136 - MultiprocessingDispatcher - 23 individuals out of 23 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:22,625 - MultiprocessingDispatcher - 11 individuals out of 11 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:25,207 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:28,801 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:36,388 - MultiprocessingDispatcher - 13 individuals out of 13 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:38,861 - MultiprocessingDispatcher - 11 individuals out of 11 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:42,152 - MultiprocessingDispatcher - 15 individuals out of 15 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:44,175 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:47,925 - MultiprocessingDispatcher - 13 individuals out of 13 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:53,183 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:55,683 - MultiprocessingDispatcher - 11 individuals out of 11 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:58,127 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 16:53:58,140 - MultiprocessingDispatcher - 19 individuals out of 19 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:03,174 - MultiprocessingDispatcher - 13 individuals out of 13 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:04,265 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:06,489 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:09,626 - MultiprocessingDispatcher - 11 individuals out of 11 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:12,062 - MultiprocessingDispatcher - 9 individuals out of 9 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:14,175 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:17,383 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:20,371 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:21,902 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:22,741 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:25,706 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:28,332 - MultiprocessingDispatcher - 11 individuals out of 11 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:28,344 - MultiprocessingDispatcher - 18 individuals out of 18 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:30,050 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:31,715 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:33,755 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 16:54:33,827 - GroupedCondition - Optimisation stopped: Time limit is reached\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generations:   0%|          | 1/10000 [09:47<?, ?gen/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 16:54:34,107 - ApiComposer - Model generation finished\n",
      "2023-10-09 16:54:34,294 - FEDOT logger - Final pipeline was fitted\n",
      "2023-10-09 16:54:34,296 - FEDOT logger - Final pipeline: {'depth': 6, 'length': 6, 'nodes': [ridge, normalization, fast_ica, normalization, pca, resample]}\n",
      "ridge - {'alpha': 6.520115566528057}\n",
      "normalization - {}\n",
      "fast_ica - {'whiten': 'unit-variance', 'n_components': 6, 'fun': 'cube'}\n",
      "normalization - {}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.7}\n",
      "resample - {'balance': 'expand_minority', 'replace': False, 'balance_ratio': 1}\n",
      "2023-10-09 16:54:34,298 - MemoryAnalytics - Memory consumption for finish in main session: current 3.6 MiB, max: 12.1 MiB\n",
      "2023-10-09 16:54:38,977 - AssumptionsHandler - Memory consumption for fitting of the initial pipeline in main session: current 0.7 MiB, max: 5.4 MiB\n",
      "2023-10-09 16:54:38,978 - ApiComposer - Initial pipeline was fitted in 1.6 sec.\n",
      "2023-10-09 16:54:38,980 - AssumptionsHandler - Preset was changed to best_quality due to fit time estimation for initial model.\n",
      "2023-10-09 16:54:38,985 - ApiComposer - AutoML configured. Parameters tuning: False. Time limit: 10 min. Set of candidate models: ['rfr', 'ridge', 'scaling', 'normalization', 'pca', 'catboostreg', 'xgbreg', 'svr', 'dtreg', 'treg', 'knnreg', 'fast_ica', 'kernel_pca', 'isolation_forest_reg', 'rfe_lin_reg', 'rfe_non_lin_reg'].\n",
      "2023-10-09 16:54:38,987 - ApiComposer - Pipeline composition started.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generations:   0%|          | 1/10000 [00:00<?, ?gen/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 16:54:40,873 - MultiprocessingDispatcher - 2 individuals out of 2 in previous population were evaluated successfully.\n",
      "2023-10-09 16:55:11,070 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 16:55:33,741 - MultiprocessingDispatcher - 12 individuals out of 12 in previous population were evaluated successfully.\n",
      "2023-10-09 16:55:36,538 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 16:55:37,357 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 16:55:50,449 - MultiprocessingDispatcher - 20 individuals out of 20 in previous population were evaluated successfully.\n",
      "2023-10-09 16:55:53,690 - MultiprocessingDispatcher - 14 individuals out of 14 in previous population were evaluated successfully.\n",
      "2023-10-09 16:56:03,136 - MultiprocessingDispatcher - 25 individuals out of 25 in previous population were evaluated successfully.\n",
      "2023-10-09 16:56:12,187 - MultiprocessingDispatcher - 24 individuals out of 25 in previous population were evaluated successfully.\n",
      "2023-10-09 16:56:13,373 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 16:56:29,271 - MultiprocessingDispatcher - 42 individuals out of 43 in previous population were evaluated successfully.\n",
      "2023-10-09 16:56:32,010 - MultiprocessingDispatcher - 11 individuals out of 11 in previous population were evaluated successfully.\n",
      "2023-10-09 16:56:32,034 - MultiprocessingDispatcher - 46 individuals out of 46 in previous population were evaluated successfully.\n",
      "2023-10-09 16:56:59,663 - MultiprocessingDispatcher - 39 individuals out of 39 in previous population were evaluated successfully.\n",
      "2023-10-09 16:57:01,124 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 16:57:07,695 - MultiprocessingDispatcher - 35 individuals out of 36 in previous population were evaluated successfully.\n",
      "2023-10-09 16:57:15,689 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 16:57:24,380 - MultiprocessingDispatcher - 34 individuals out of 36 in previous population were evaluated successfully.\n",
      "2023-10-09 16:57:31,214 - MultiprocessingDispatcher - 32 individuals out of 33 in previous population were evaluated successfully.\n",
      "2023-10-09 16:57:37,712 - MultiprocessingDispatcher - 29 individuals out of 29 in previous population were evaluated successfully.\n",
      "2023-10-09 16:57:42,575 - MultiprocessingDispatcher - 30 individuals out of 30 in previous population were evaluated successfully.\n",
      "2023-10-09 16:57:42,599 - MultiprocessingDispatcher - 51 individuals out of 51 in previous population were evaluated successfully.\n",
      "2023-10-09 16:57:53,732 - MultiprocessingDispatcher - 42 individuals out of 42 in previous population were evaluated successfully.\n",
      "2023-10-09 16:57:55,956 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 16:58:18,467 - MultiprocessingDispatcher - 41 individuals out of 41 in previous population were evaluated successfully.\n",
      "2023-10-09 16:58:21,383 - MultiprocessingDispatcher - 9 individuals out of 9 in previous population were evaluated successfully.\n",
      "2023-10-09 16:58:40,283 - MultiprocessingDispatcher - 42 individuals out of 43 in previous population were evaluated successfully.\n",
      "2023-10-09 16:58:41,911 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 16:58:55,368 - MultiprocessingDispatcher - 42 individuals out of 42 in previous population were evaluated successfully.\n",
      "2023-10-09 16:58:57,820 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 16:59:16,115 - MultiprocessingDispatcher - 44 individuals out of 44 in previous population were evaluated successfully.\n",
      "2023-10-09 16:59:18,095 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 16:59:18,120 - MultiprocessingDispatcher - 41 individuals out of 41 in previous population were evaluated successfully.\n",
      "2023-10-09 16:59:29,654 - MultiprocessingDispatcher - 35 individuals out of 36 in previous population were evaluated successfully.\n",
      "2023-10-09 16:59:38,605 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 16:59:59,135 - MultiprocessingDispatcher - 34 individuals out of 34 in previous population were evaluated successfully.\n",
      "2023-10-09 17:00:00,724 - MultiprocessingDispatcher - 7 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 17:00:30,710 - MultiprocessingDispatcher - 36 individuals out of 36 in previous population were evaluated successfully.\n",
      "2023-10-09 17:00:32,441 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 17:00:40,869 - MultiprocessingDispatcher - 32 individuals out of 32 in previous population were evaluated successfully.\n",
      "2023-10-09 17:00:43,383 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 17:01:08,351 - MultiprocessingDispatcher - 36 individuals out of 36 in previous population were evaluated successfully.\n",
      "2023-10-09 17:01:10,303 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 17:01:10,328 - MultiprocessingDispatcher - 36 individuals out of 36 in previous population were evaluated successfully.\n",
      "2023-10-09 17:01:17,999 - MultiprocessingDispatcher - 32 individuals out of 33 in previous population were evaluated successfully.\n",
      "2023-10-09 17:01:19,420 - MultiprocessingDispatcher - 2 individuals out of 2 in previous population were evaluated successfully.\n",
      "2023-10-09 17:01:28,681 - MultiprocessingDispatcher - 32 individuals out of 32 in previous population were evaluated successfully.\n",
      "2023-10-09 17:01:30,575 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:01:39,077 - MultiprocessingDispatcher - 30 individuals out of 31 in previous population were evaluated successfully.\n",
      "2023-10-09 17:01:40,973 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 17:02:05,984 - MultiprocessingDispatcher - 32 individuals out of 32 in previous population were evaluated successfully.\n",
      "2023-10-09 17:02:07,325 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:02:17,666 - MultiprocessingDispatcher - 29 individuals out of 29 in previous population were evaluated successfully.\n",
      "2023-10-09 17:02:19,654 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 17:02:19,673 - MultiprocessingDispatcher - 34 individuals out of 34 in previous population were evaluated successfully.\n",
      "2023-10-09 17:02:31,486 - MultiprocessingDispatcher - 33 individuals out of 33 in previous population were evaluated successfully.\n",
      "2023-10-09 17:02:47,835 - MultiprocessingDispatcher - 29 individuals out of 29 in previous population were evaluated successfully.\n",
      "2023-10-09 17:02:49,400 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:02:57,576 - MultiprocessingDispatcher - 30 individuals out of 30 in previous population were evaluated successfully.\n",
      "2023-10-09 17:03:02,763 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:03:09,865 - MultiprocessingDispatcher - 24 individuals out of 24 in previous population were evaluated successfully.\n",
      "2023-10-09 17:03:12,960 - MultiprocessingDispatcher - 9 individuals out of 9 in previous population were evaluated successfully.\n",
      "2023-10-09 17:03:20,176 - MultiprocessingDispatcher - 27 individuals out of 27 in previous population were evaluated successfully.\n",
      "2023-10-09 17:03:23,043 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 17:03:23,065 - MultiprocessingDispatcher - 29 individuals out of 29 in previous population were evaluated successfully.\n",
      "2023-10-09 17:03:31,423 - MultiprocessingDispatcher - 25 individuals out of 25 in previous population were evaluated successfully.\n",
      "2023-10-09 17:03:32,381 - MultiprocessingDispatcher - 2 individuals out of 2 in previous population were evaluated successfully.\n",
      "2023-10-09 17:03:45,745 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 17:03:47,246 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 17:03:53,404 - MultiprocessingDispatcher - 24 individuals out of 24 in previous population were evaluated successfully.\n",
      "2023-10-09 17:03:55,495 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 17:04:02,949 - MultiprocessingDispatcher - 23 individuals out of 23 in previous population were evaluated successfully.\n",
      "2023-10-09 17:04:07,106 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 17:04:16,718 - MultiprocessingDispatcher - 23 individuals out of 23 in previous population were evaluated successfully.\n",
      "2023-10-09 17:04:18,679 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 17:04:18,696 - MultiprocessingDispatcher - 26 individuals out of 26 in previous population were evaluated successfully.\n",
      "2023-10-09 17:04:31,815 - MultiprocessingDispatcher - 23 individuals out of 23 in previous population were evaluated successfully.\n",
      "2023-10-09 17:04:33,865 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:04:33,965 - GroupedCondition - Optimisation stopped: Time limit is reached\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generations:   0%|          | 1/10000 [09:54<?, ?gen/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 17:04:34,142 - ApiComposer - Model generation finished\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 17:04:35,088 - FEDOT logger - Final pipeline was fitted\n",
      "2023-10-09 17:04:35,090 - FEDOT logger - Final pipeline: {'depth': 6, 'length': 8, 'nodes': [ridge, pca, isolation_forest_reg, resample, pca, normalization, normalization, scaling]}\n",
      "ridge - {'alpha': 4.4996825838125245}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.7}\n",
      "isolation_forest_reg - {'max_samples': 0.8784623773028627, 'max_features': 0.4819803289315286, 'bootstrap': False}\n",
      "resample - {'balance': 'expand_minority', 'replace': False, 'balance_ratio': 0.9345135904539794}\n",
      "pca - {'svd_solver': 'full', 'n_components': 0.7}\n",
      "normalization - {}\n",
      "normalization - {}\n",
      "scaling - {}\n",
      "2023-10-09 17:04:35,091 - MemoryAnalytics - Memory consumption for finish in main session: current 4.6 MiB, max: 12.1 MiB\n",
      "2023-10-09 17:04:39,766 - AssumptionsHandler - Memory consumption for fitting of the initial pipeline in main session: current 0.7 MiB, max: 5.4 MiB\n",
      "2023-10-09 17:04:39,768 - ApiComposer - Initial pipeline was fitted in 1.5 sec.\n",
      "2023-10-09 17:04:39,769 - AssumptionsHandler - Preset was changed to best_quality due to fit time estimation for initial model.\n",
      "2023-10-09 17:04:39,775 - ApiComposer - AutoML configured. Parameters tuning: False. Time limit: 10 min. Set of candidate models: ['rfr', 'ridge', 'scaling', 'normalization', 'pca', 'catboostreg', 'xgbreg', 'svr', 'dtreg', 'treg', 'knnreg', 'fast_ica', 'kernel_pca', 'isolation_forest_reg', 'rfe_lin_reg', 'rfe_non_lin_reg'].\n",
      "2023-10-09 17:04:39,777 - ApiComposer - Pipeline composition started.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generations:   0%|          | 1/10000 [00:00<?, ?gen/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 17:04:41,352 - MultiprocessingDispatcher - 2 individuals out of 2 in previous population were evaluated successfully.\n",
      "2023-10-09 17:05:06,168 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 17:05:38,540 - MultiprocessingDispatcher - 16 individuals out of 16 in previous population were evaluated successfully.\n",
      "2023-10-09 17:05:39,693 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:06:21,302 - MultiprocessingDispatcher - 16 individuals out of 16 in previous population were evaluated successfully.\n",
      "2023-10-09 17:06:23,383 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:06:28,678 - MultiprocessingDispatcher - 20 individuals out of 20 in previous population were evaluated successfully.\n",
      "2023-10-09 17:06:38,770 - MultiprocessingDispatcher - 14 individuals out of 14 in previous population were evaluated successfully.\n",
      "2023-10-09 17:06:54,835 - MultiprocessingDispatcher - 29 individuals out of 29 in previous population were evaluated successfully.\n",
      "2023-10-09 17:07:04,668 - MultiprocessingDispatcher - 22 individuals out of 22 in previous population were evaluated successfully.\n",
      "2023-10-09 17:07:04,688 - MultiprocessingDispatcher - 41 individuals out of 41 in previous population were evaluated successfully.\n",
      "2023-10-09 17:07:16,433 - MultiprocessingDispatcher - 31 individuals out of 31 in previous population were evaluated successfully.\n",
      "2023-10-09 17:07:19,360 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 17:07:34,093 - MultiprocessingDispatcher - 33 individuals out of 34 in previous population were evaluated successfully.\n",
      "2023-10-09 17:07:41,951 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 17:07:56,914 - MultiprocessingDispatcher - 31 individuals out of 31 in previous population were evaluated successfully.\n",
      "2023-10-09 17:07:59,992 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 17:08:14,936 - MultiprocessingDispatcher - 32 individuals out of 33 in previous population were evaluated successfully.\n",
      "2023-10-09 17:08:16,465 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 17:08:35,046 - MultiprocessingDispatcher - 36 individuals out of 36 in previous population were evaluated successfully.\n",
      "2023-10-09 17:08:36,191 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 17:08:36,213 - MultiprocessingDispatcher - 35 individuals out of 35 in previous population were evaluated successfully.\n",
      "2023-10-09 17:08:47,616 - MultiprocessingDispatcher - 27 individuals out of 27 in previous population were evaluated successfully.\n",
      "2023-10-09 17:08:50,059 - MultiprocessingDispatcher - 9 individuals out of 9 in previous population were evaluated successfully.\n",
      "2023-10-09 17:09:01,866 - MultiprocessingDispatcher - 30 individuals out of 30 in previous population were evaluated successfully.\n",
      "2023-10-09 17:09:03,386 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 17:09:09,453 - MultiprocessingDispatcher - 28 individuals out of 29 in previous population were evaluated successfully.\n",
      "2023-10-09 17:09:19,988 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 17:09:30,268 - MultiprocessingDispatcher - 29 individuals out of 31 in previous population were evaluated successfully.\n",
      "2023-10-09 17:09:32,168 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 17:09:55,688 - MultiprocessingDispatcher - 31 individuals out of 31 in previous population were evaluated successfully.\n",
      "2023-10-09 17:10:00,119 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:10:00,134 - MultiprocessingDispatcher - 33 individuals out of 33 in previous population were evaluated successfully.\n",
      "2023-10-09 17:10:07,691 - MultiprocessingDispatcher - 25 individuals out of 26 in previous population were evaluated successfully.\n",
      "2023-10-09 17:10:09,557 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 17:10:18,905 - MultiprocessingDispatcher - 26 individuals out of 26 in previous population were evaluated successfully.\n",
      "2023-10-09 17:10:25,570 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 17:10:36,307 - MultiprocessingDispatcher - 30 individuals out of 30 in previous population were evaluated successfully.\n",
      "2023-10-09 17:10:43,781 - MultiprocessingDispatcher - 28 individuals out of 28 in previous population were evaluated successfully.\n",
      "2023-10-09 17:10:45,079 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:10:53,717 - MultiprocessingDispatcher - 30 individuals out of 30 in previous population were evaluated successfully.\n",
      "2023-10-09 17:10:53,731 - MultiprocessingDispatcher - 25 individuals out of 25 in previous population were evaluated successfully.\n",
      "2023-10-09 17:10:59,714 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:00,902 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:06,875 - MultiprocessingDispatcher - 20 individuals out of 20 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:08,235 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:14,564 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:15,503 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:22,260 - MultiprocessingDispatcher - 20 individuals out of 20 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:25,931 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:30,363 - MultiprocessingDispatcher - 18 individuals out of 18 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:31,468 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:31,480 - MultiprocessingDispatcher - 21 individuals out of 21 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:37,487 - MultiprocessingDispatcher - 16 individuals out of 16 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:43,731 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:53,723 - MultiprocessingDispatcher - 17 individuals out of 17 in previous population were evaluated successfully.\n",
      "2023-10-09 17:11:54,847 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:03,493 - MultiprocessingDispatcher - 19 individuals out of 19 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:11,526 - MultiprocessingDispatcher - 18 individuals out of 18 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:14,885 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:18,748 - MultiprocessingDispatcher - 16 individuals out of 16 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:20,324 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:20,333 - MultiprocessingDispatcher - 19 individuals out of 19 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:23,870 - MultiprocessingDispatcher - 15 individuals out of 15 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:25,628 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:34,109 - MultiprocessingDispatcher - 15 individuals out of 15 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:34,889 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:41,512 - MultiprocessingDispatcher - 16 individuals out of 16 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:42,740 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:48,493 - MultiprocessingDispatcher - 12 individuals out of 12 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:53,440 - MultiprocessingDispatcher - 7 individuals out of 7 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:56,480 - MultiprocessingDispatcher - 16 individuals out of 16 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:57,210 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:57,219 - MultiprocessingDispatcher - 14 individuals out of 14 in previous population were evaluated successfully.\n",
      "2023-10-09 17:12:59,268 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:02,668 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:05,404 - MultiprocessingDispatcher - 13 individuals out of 13 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:08,869 - MultiprocessingDispatcher - 14 individuals out of 14 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:13,071 - MultiprocessingDispatcher - 14 individuals out of 14 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:15,242 - MultiprocessingDispatcher - 12 individuals out of 12 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:16,086 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:16,092 - MultiprocessingDispatcher - 12 individuals out of 12 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:17,759 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:19,040 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:20,827 - MultiprocessingDispatcher - 9 individuals out of 9 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:22,289 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:24,391 - MultiprocessingDispatcher - 6 individuals out of 6 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:27,003 - MultiprocessingDispatcher - 5 individuals out of 5 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:34,811 - MultiprocessingDispatcher - 12 individuals out of 12 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:40,638 - MultiprocessingDispatcher - 12 individuals out of 12 in previous population were evaluated successfully.\n",
      "2023-10-09 17:13:40,644 - MultiprocessingDispatcher - 11 individuals out of 11 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:00,446 - MultiprocessingDispatcher - 9 individuals out of 9 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:01,922 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:07,258 - MultiprocessingDispatcher - 9 individuals out of 9 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:08,070 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:11,181 - MultiprocessingDispatcher - 8 individuals out of 8 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:12,757 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:15,074 - MultiprocessingDispatcher - 9 individuals out of 9 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:16,430 - MultiprocessingDispatcher - 4 individuals out of 4 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:21,583 - MultiprocessingDispatcher - 9 individuals out of 9 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:22,805 - MultiprocessingDispatcher - 3 individuals out of 3 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:22,811 - MultiprocessingDispatcher - 11 individuals out of 11 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:28,385 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:32,689 - MultiprocessingDispatcher - 10 individuals out of 10 in previous population were evaluated successfully.\n",
      "2023-10-09 17:14:32,722 - GroupedCondition - Optimisation stopped: Time limit is reached\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Generations:   0%|          | 1/10000 [09:52<?, ?gen/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2023-10-09 17:14:32,970 - ApiComposer - Model generation finished\n",
      "2023-10-09 17:14:33,875 - FEDOT logger - Final pipeline was fitted\n",
      "2023-10-09 17:14:33,877 - FEDOT logger - Final pipeline: {'depth': 3, 'length': 3, 'nodes': [ridge, isolation_forest_reg, normalization]}\n",
      "ridge - {'alpha': 2.8230574623414832}\n",
      "isolation_forest_reg - {'max_samples': 0.8564856493691568, 'max_features': 0.08359953031874037, 'bootstrap': False}\n",
      "normalization - {}\n",
      "2023-10-09 17:14:33,879 - MemoryAnalytics - Memory consumption for finish in main session: current 3.1 MiB, max: 7.9 MiB\n"
     ]
    }
   ],
   "source": [
    "metric_dict = {}\n",
    "runs = 5\n",
    "for run in range(runs):\n",
    "    with IndustrialModels():\n",
    "        auto_ml = PipelineBuilder().add_node('recurrence_extractor',\n",
    "                                              params={'window_size': 25, 'stride': 2}).add_node('fedot_regr',\n",
    "                                                                             params={'timeout': 10,'logging_level':30}).build()\n",
    "        auto_ml.fit(input_data)\n",
    "        features = auto_ml.predict(val_data).predict\n",
    "        metric = calculate_regression_metric(test_target=test_data[1], labels=features)\n",
    "        metric = metric.T\n",
    "        metric.columns = metric.columns.values\n",
    "        metric['model_params'] = metric['model_params'] = str({node:node.parameters for node in auto_ml.graph_description['nodes']})\n",
    "        metric_dict.update({f'run_number - {run}':metric})"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "outputs": [
    {
     "data": {
      "text/plain": "               r2_score: mean_squared_error: root_mean_squared_error:  \\\nrun_number - 3  0.112933       234275.486586                484.02013   \nrun_number - 4   0.11024       234986.646191               484.754212   \nrun_number - 1  0.102146       237124.182285               486.953984   \nrun_number - 2  0.100784       237483.895586               487.323194   \nrun_number - 0  0.094814        239060.56708               488.938204   \n\n               mean_absolute_error median_absolute_error  \\\nrun_number - 3          248.559733            125.777939   \nrun_number - 4          242.573667            113.849569   \nrun_number - 1          240.224593            106.795772   \nrun_number - 2          246.959405            115.911299   \nrun_number - 0           234.62664            108.245939   \n\n               explained_variance_score    max_error d2_absolute_error_score  \\\nrun_number - 3                 0.116121  2155.903861               -0.288602   \nrun_number - 4                 0.116248  2127.662682               -0.257568   \nrun_number - 1                 0.108916  2140.707903                -0.24539   \nrun_number - 2                   0.1046  2086.129884               -0.280305   \nrun_number - 0                  0.09879  2199.298936               -0.216369   \n\n                                                     model_params  \nrun_number - 3  {fedot_regr: {'timeout': 10, 'logging_level': ...  \nrun_number - 4  {fedot_regr: {'timeout': 10, 'logging_level': ...  \nrun_number - 1  {fedot_regr: {'timeout': 10, 'logging_level': ...  \nrun_number - 2  {fedot_regr: {'timeout': 10, 'logging_level': ...  \nrun_number - 0  {fedot_regr: {'timeout': 10, 'logging_level': ...  ",
      "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>r2_score:</th>\n      <th>mean_squared_error:</th>\n      <th>root_mean_squared_error:</th>\n      <th>mean_absolute_error</th>\n      <th>median_absolute_error</th>\n      <th>explained_variance_score</th>\n      <th>max_error</th>\n      <th>d2_absolute_error_score</th>\n      <th>model_params</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>run_number - 3</th>\n      <td>0.112933</td>\n      <td>234275.486586</td>\n      <td>484.02013</td>\n      <td>248.559733</td>\n      <td>125.777939</td>\n      <td>0.116121</td>\n      <td>2155.903861</td>\n      <td>-0.288602</td>\n      <td>{fedot_regr: {'timeout': 10, 'logging_level': ...</td>\n    </tr>\n    <tr>\n      <th>run_number - 4</th>\n      <td>0.11024</td>\n      <td>234986.646191</td>\n      <td>484.754212</td>\n      <td>242.573667</td>\n      <td>113.849569</td>\n      <td>0.116248</td>\n      <td>2127.662682</td>\n      <td>-0.257568</td>\n      <td>{fedot_regr: {'timeout': 10, 'logging_level': ...</td>\n    </tr>\n    <tr>\n      <th>run_number - 1</th>\n      <td>0.102146</td>\n      <td>237124.182285</td>\n      <td>486.953984</td>\n      <td>240.224593</td>\n      <td>106.795772</td>\n      <td>0.108916</td>\n      <td>2140.707903</td>\n      <td>-0.24539</td>\n      <td>{fedot_regr: {'timeout': 10, 'logging_level': ...</td>\n    </tr>\n    <tr>\n      <th>run_number - 2</th>\n      <td>0.100784</td>\n      <td>237483.895586</td>\n      <td>487.323194</td>\n      <td>246.959405</td>\n      <td>115.911299</td>\n      <td>0.1046</td>\n      <td>2086.129884</td>\n      <td>-0.280305</td>\n      <td>{fedot_regr: {'timeout': 10, 'logging_level': ...</td>\n    </tr>\n    <tr>\n      <th>run_number - 0</th>\n      <td>0.094814</td>\n      <td>239060.56708</td>\n      <td>488.938204</td>\n      <td>234.62664</td>\n      <td>108.245939</td>\n      <td>0.09879</td>\n      <td>2199.298936</td>\n      <td>-0.216369</td>\n      <td>{fedot_regr: {'timeout': 10, 'logging_level': ...</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_automl = pd.concat([x for x in metric_dict.values()],axis=0).T\n",
    "df_automl.columns = list(metric_dict.keys())\n",
    "df_automl = df_automl.T\n",
    "df_automl.sort_values('root_mean_squared_error:')"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "markdown",
   "source": [
    "## Compare with State of Art (SOTA) models"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "outputs": [],
   "source": [
    "df = pd.read_csv(data_path+'/ts_regression_sota_results.csv',sep=';')"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "outputs": [],
   "source": [
    "df = df[df['ds/type'] == dataset_name].iloc[:,:25]\n",
    "df.index = df['algorithm']\n",
    "df = df.drop(['algorithm','ds/type'], axis=1)\n",
    "df = df.replace(',','.', regex=True).astype(float)\n",
    "df['Fedot_Industrial_baseline'] = best_baseline\n",
    "df['Fedot_Industrial_tuned'] = best_tuned\n",
    "df['Fedot_Industrial_AutoML'] = 0\n",
    "df['Fedot_Industrial_AutoML'].iloc[0] = df_automl['root_mean_squared_error:'].min()\n",
    "df['Fedot_Industrial_AutoML'].iloc[1] = df_automl['root_mean_squared_error:'].max()\n",
    "df['Fedot_Industrial_AutoML'].iloc[2] = df_automl['root_mean_squared_error:'].mean()\n",
    "df = df.T"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "outputs": [
    {
     "data": {
      "text/plain": "algorithm                         min         max     average\nCNN_RMSE                   387.616535  561.795243  484.637414\nGrid-SVR_RMSE              390.641257  553.666929  464.389249\nFPCR_RMSE                  399.459870  522.998212  455.941085\n5NN-DTW_RMSE               407.790548  546.981708  469.378018\nFPCR-Bs_RMSE               408.199670  537.381895  465.844347\n5NN-ED_RMSE                424.504623  562.413046  493.478164\nRandF_RMSE                 456.359415  589.478468  527.343198\nFedot_Industrial_AutoML    484.020130  488.938204  486.397945\nFreshPRINCE_RMSE           487.642605  640.463800  582.654848\nResNet_RMSE                489.719817  661.954851  589.427954\nRDST_RMSE                  491.072419  575.720140  527.927002\nFedot_Industrial_tuned     491.323216  491.323216  491.323216\nFedot_Industrial_baseline  492.792531  492.792531  492.792531\nROCKET_RMSE                492.914183  629.215008  575.695827\nDrCIF_RMSE                 494.923063  646.879703  587.404402\nFCN_RMSE                   499.100408  693.068654  609.970047\nSingleInception_RMSE       499.449452  652.585529  588.659847\nInceptionT_RMSE            506.696071  645.057362  589.468140\nMultiROCKET_RMSE           506.770267  663.968054  596.389600\nRidge_RMSE                 509.119169  627.513473  574.558434\nTSF_RMSE                   510.156025  663.258974  604.193543\nRotF_RMSE                  519.799244  670.442108  606.662679\nXGBoost_RMSE               526.421564  693.574841  618.172663\n1NN-ED_RMSE                529.096971  812.146002  684.484545\n1NN-DTW_RMSE               557.562623  755.854409  666.809599\nRIST_RMSE                  560.257921  639.787936  601.625863",
      "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th>algorithm</th>\n      <th>min</th>\n      <th>max</th>\n      <th>average</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>CNN_RMSE</th>\n      <td>387.616535</td>\n      <td>561.795243</td>\n      <td>484.637414</td>\n    </tr>\n    <tr>\n      <th>Grid-SVR_RMSE</th>\n      <td>390.641257</td>\n      <td>553.666929</td>\n      <td>464.389249</td>\n    </tr>\n    <tr>\n      <th>FPCR_RMSE</th>\n      <td>399.459870</td>\n      <td>522.998212</td>\n      <td>455.941085</td>\n    </tr>\n    <tr>\n      <th>5NN-DTW_RMSE</th>\n      <td>407.790548</td>\n      <td>546.981708</td>\n      <td>469.378018</td>\n    </tr>\n    <tr>\n      <th>FPCR-Bs_RMSE</th>\n      <td>408.199670</td>\n      <td>537.381895</td>\n      <td>465.844347</td>\n    </tr>\n    <tr>\n      <th>5NN-ED_RMSE</th>\n      <td>424.504623</td>\n      <td>562.413046</td>\n      <td>493.478164</td>\n    </tr>\n    <tr>\n      <th>RandF_RMSE</th>\n      <td>456.359415</td>\n      <td>589.478468</td>\n      <td>527.343198</td>\n    </tr>\n    <tr>\n      <th>Fedot_Industrial_AutoML</th>\n      <td>484.020130</td>\n      <td>488.938204</td>\n      <td>486.397945</td>\n    </tr>\n    <tr>\n      <th>FreshPRINCE_RMSE</th>\n      <td>487.642605</td>\n      <td>640.463800</td>\n      <td>582.654848</td>\n    </tr>\n    <tr>\n      <th>ResNet_RMSE</th>\n      <td>489.719817</td>\n      <td>661.954851</td>\n      <td>589.427954</td>\n    </tr>\n    <tr>\n      <th>RDST_RMSE</th>\n      <td>491.072419</td>\n      <td>575.720140</td>\n      <td>527.927002</td>\n    </tr>\n    <tr>\n      <th>Fedot_Industrial_tuned</th>\n      <td>491.323216</td>\n      <td>491.323216</td>\n      <td>491.323216</td>\n    </tr>\n    <tr>\n      <th>Fedot_Industrial_baseline</th>\n      <td>492.792531</td>\n      <td>492.792531</td>\n      <td>492.792531</td>\n    </tr>\n    <tr>\n      <th>ROCKET_RMSE</th>\n      <td>492.914183</td>\n      <td>629.215008</td>\n      <td>575.695827</td>\n    </tr>\n    <tr>\n      <th>DrCIF_RMSE</th>\n      <td>494.923063</td>\n      <td>646.879703</td>\n      <td>587.404402</td>\n    </tr>\n    <tr>\n      <th>FCN_RMSE</th>\n      <td>499.100408</td>\n      <td>693.068654</td>\n      <td>609.970047</td>\n    </tr>\n    <tr>\n      <th>SingleInception_RMSE</th>\n      <td>499.449452</td>\n      <td>652.585529</td>\n      <td>588.659847</td>\n    </tr>\n    <tr>\n      <th>InceptionT_RMSE</th>\n      <td>506.696071</td>\n      <td>645.057362</td>\n      <td>589.468140</td>\n    </tr>\n    <tr>\n      <th>MultiROCKET_RMSE</th>\n      <td>506.770267</td>\n      <td>663.968054</td>\n      <td>596.389600</td>\n    </tr>\n    <tr>\n      <th>Ridge_RMSE</th>\n      <td>509.119169</td>\n      <td>627.513473</td>\n      <td>574.558434</td>\n    </tr>\n    <tr>\n      <th>TSF_RMSE</th>\n      <td>510.156025</td>\n      <td>663.258974</td>\n      <td>604.193543</td>\n    </tr>\n    <tr>\n      <th>RotF_RMSE</th>\n      <td>519.799244</td>\n      <td>670.442108</td>\n      <td>606.662679</td>\n    </tr>\n    <tr>\n      <th>XGBoost_RMSE</th>\n      <td>526.421564</td>\n      <td>693.574841</td>\n      <td>618.172663</td>\n    </tr>\n    <tr>\n      <th>1NN-ED_RMSE</th>\n      <td>529.096971</td>\n      <td>812.146002</td>\n      <td>684.484545</td>\n    </tr>\n    <tr>\n      <th>1NN-DTW_RMSE</th>\n      <td>557.562623</td>\n      <td>755.854409</td>\n      <td>666.809599</td>\n    </tr>\n    <tr>\n      <th>RIST_RMSE</th>\n      <td>560.257921</td>\n      <td>639.787936</td>\n      <td>601.625863</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sort_values('min')"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "outputs": [
    {
     "data": {
      "text/plain": "algorithm                         min         max     average\nFedot_Industrial_AutoML    484.020130  488.938204  486.397945\nFedot_Industrial_tuned     491.323216  491.323216  491.323216\nFedot_Industrial_baseline  492.792531  492.792531  492.792531\nFPCR_RMSE                  399.459870  522.998212  455.941085\nFPCR-Bs_RMSE               408.199670  537.381895  465.844347\n5NN-DTW_RMSE               407.790548  546.981708  469.378018\nGrid-SVR_RMSE              390.641257  553.666929  464.389249\nCNN_RMSE                   387.616535  561.795243  484.637414\n5NN-ED_RMSE                424.504623  562.413046  493.478164\nRDST_RMSE                  491.072419  575.720140  527.927002\nRandF_RMSE                 456.359415  589.478468  527.343198\nRidge_RMSE                 509.119169  627.513473  574.558434\nROCKET_RMSE                492.914183  629.215008  575.695827\nRIST_RMSE                  560.257921  639.787936  601.625863\nFreshPRINCE_RMSE           487.642605  640.463800  582.654848\nInceptionT_RMSE            506.696071  645.057362  589.468140\nDrCIF_RMSE                 494.923063  646.879703  587.404402\nSingleInception_RMSE       499.449452  652.585529  588.659847\nResNet_RMSE                489.719817  661.954851  589.427954\nTSF_RMSE                   510.156025  663.258974  604.193543\nMultiROCKET_RMSE           506.770267  663.968054  596.389600\nRotF_RMSE                  519.799244  670.442108  606.662679\nFCN_RMSE                   499.100408  693.068654  609.970047\nXGBoost_RMSE               526.421564  693.574841  618.172663\n1NN-DTW_RMSE               557.562623  755.854409  666.809599\n1NN-ED_RMSE                529.096971  812.146002  684.484545",
      "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th>algorithm</th>\n      <th>min</th>\n      <th>max</th>\n      <th>average</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>Fedot_Industrial_AutoML</th>\n      <td>484.020130</td>\n      <td>488.938204</td>\n      <td>486.397945</td>\n    </tr>\n    <tr>\n      <th>Fedot_Industrial_tuned</th>\n      <td>491.323216</td>\n      <td>491.323216</td>\n      <td>491.323216</td>\n    </tr>\n    <tr>\n      <th>Fedot_Industrial_baseline</th>\n      <td>492.792531</td>\n      <td>492.792531</td>\n      <td>492.792531</td>\n    </tr>\n    <tr>\n      <th>FPCR_RMSE</th>\n      <td>399.459870</td>\n      <td>522.998212</td>\n      <td>455.941085</td>\n    </tr>\n    <tr>\n      <th>FPCR-Bs_RMSE</th>\n      <td>408.199670</td>\n      <td>537.381895</td>\n      <td>465.844347</td>\n    </tr>\n    <tr>\n      <th>5NN-DTW_RMSE</th>\n      <td>407.790548</td>\n      <td>546.981708</td>\n      <td>469.378018</td>\n    </tr>\n    <tr>\n      <th>Grid-SVR_RMSE</th>\n      <td>390.641257</td>\n      <td>553.666929</td>\n      <td>464.389249</td>\n    </tr>\n    <tr>\n      <th>CNN_RMSE</th>\n      <td>387.616535</td>\n      <td>561.795243</td>\n      <td>484.637414</td>\n    </tr>\n    <tr>\n      <th>5NN-ED_RMSE</th>\n      <td>424.504623</td>\n      <td>562.413046</td>\n      <td>493.478164</td>\n    </tr>\n    <tr>\n      <th>RDST_RMSE</th>\n      <td>491.072419</td>\n      <td>575.720140</td>\n      <td>527.927002</td>\n    </tr>\n    <tr>\n      <th>RandF_RMSE</th>\n      <td>456.359415</td>\n      <td>589.478468</td>\n      <td>527.343198</td>\n    </tr>\n    <tr>\n      <th>Ridge_RMSE</th>\n      <td>509.119169</td>\n      <td>627.513473</td>\n      <td>574.558434</td>\n    </tr>\n    <tr>\n      <th>ROCKET_RMSE</th>\n      <td>492.914183</td>\n      <td>629.215008</td>\n      <td>575.695827</td>\n    </tr>\n    <tr>\n      <th>RIST_RMSE</th>\n      <td>560.257921</td>\n      <td>639.787936</td>\n      <td>601.625863</td>\n    </tr>\n    <tr>\n      <th>FreshPRINCE_RMSE</th>\n      <td>487.642605</td>\n      <td>640.463800</td>\n      <td>582.654848</td>\n    </tr>\n    <tr>\n      <th>InceptionT_RMSE</th>\n      <td>506.696071</td>\n      <td>645.057362</td>\n      <td>589.468140</td>\n    </tr>\n    <tr>\n      <th>DrCIF_RMSE</th>\n      <td>494.923063</td>\n      <td>646.879703</td>\n      <td>587.404402</td>\n    </tr>\n    <tr>\n      <th>SingleInception_RMSE</th>\n      <td>499.449452</td>\n      <td>652.585529</td>\n      <td>588.659847</td>\n    </tr>\n    <tr>\n      <th>ResNet_RMSE</th>\n      <td>489.719817</td>\n      <td>661.954851</td>\n      <td>589.427954</td>\n    </tr>\n    <tr>\n      <th>TSF_RMSE</th>\n      <td>510.156025</td>\n      <td>663.258974</td>\n      <td>604.193543</td>\n    </tr>\n    <tr>\n      <th>MultiROCKET_RMSE</th>\n      <td>506.770267</td>\n      <td>663.968054</td>\n      <td>596.389600</td>\n    </tr>\n    <tr>\n      <th>RotF_RMSE</th>\n      <td>519.799244</td>\n      <td>670.442108</td>\n      <td>606.662679</td>\n    </tr>\n    <tr>\n      <th>FCN_RMSE</th>\n      <td>499.100408</td>\n      <td>693.068654</td>\n      <td>609.970047</td>\n    </tr>\n    <tr>\n      <th>XGBoost_RMSE</th>\n      <td>526.421564</td>\n      <td>693.574841</td>\n      <td>618.172663</td>\n    </tr>\n    <tr>\n      <th>1NN-DTW_RMSE</th>\n      <td>557.562623</td>\n      <td>755.854409</td>\n      <td>666.809599</td>\n    </tr>\n    <tr>\n      <th>1NN-ED_RMSE</th>\n      <td>529.096971</td>\n      <td>812.146002</td>\n      <td>684.484545</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sort_values('max')"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "outputs": [
    {
     "data": {
      "text/plain": "algorithm                         min         max     average\nFPCR_RMSE                  399.459870  522.998212  455.941085\nGrid-SVR_RMSE              390.641257  553.666929  464.389249\nFPCR-Bs_RMSE               408.199670  537.381895  465.844347\n5NN-DTW_RMSE               407.790548  546.981708  469.378018\nCNN_RMSE                   387.616535  561.795243  484.637414\nFedot_Industrial_AutoML    484.020130  488.938204  486.397945\nFedot_Industrial_tuned     491.323216  491.323216  491.323216\nFedot_Industrial_baseline  492.792531  492.792531  492.792531\n5NN-ED_RMSE                424.504623  562.413046  493.478164\nRandF_RMSE                 456.359415  589.478468  527.343198\nRDST_RMSE                  491.072419  575.720140  527.927002\nRidge_RMSE                 509.119169  627.513473  574.558434\nROCKET_RMSE                492.914183  629.215008  575.695827\nFreshPRINCE_RMSE           487.642605  640.463800  582.654848\nDrCIF_RMSE                 494.923063  646.879703  587.404402\nSingleInception_RMSE       499.449452  652.585529  588.659847\nResNet_RMSE                489.719817  661.954851  589.427954\nInceptionT_RMSE            506.696071  645.057362  589.468140\nMultiROCKET_RMSE           506.770267  663.968054  596.389600\nRIST_RMSE                  560.257921  639.787936  601.625863\nTSF_RMSE                   510.156025  663.258974  604.193543\nRotF_RMSE                  519.799244  670.442108  606.662679\nFCN_RMSE                   499.100408  693.068654  609.970047\nXGBoost_RMSE               526.421564  693.574841  618.172663\n1NN-DTW_RMSE               557.562623  755.854409  666.809599\n1NN-ED_RMSE                529.096971  812.146002  684.484545",
      "text/html": "<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th>algorithm</th>\n      <th>min</th>\n      <th>max</th>\n      <th>average</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>FPCR_RMSE</th>\n      <td>399.459870</td>\n      <td>522.998212</td>\n      <td>455.941085</td>\n    </tr>\n    <tr>\n      <th>Grid-SVR_RMSE</th>\n      <td>390.641257</td>\n      <td>553.666929</td>\n      <td>464.389249</td>\n    </tr>\n    <tr>\n      <th>FPCR-Bs_RMSE</th>\n      <td>408.199670</td>\n      <td>537.381895</td>\n      <td>465.844347</td>\n    </tr>\n    <tr>\n      <th>5NN-DTW_RMSE</th>\n      <td>407.790548</td>\n      <td>546.981708</td>\n      <td>469.378018</td>\n    </tr>\n    <tr>\n      <th>CNN_RMSE</th>\n      <td>387.616535</td>\n      <td>561.795243</td>\n      <td>484.637414</td>\n    </tr>\n    <tr>\n      <th>Fedot_Industrial_AutoML</th>\n      <td>484.020130</td>\n      <td>488.938204</td>\n      <td>486.397945</td>\n    </tr>\n    <tr>\n      <th>Fedot_Industrial_tuned</th>\n      <td>491.323216</td>\n      <td>491.323216</td>\n      <td>491.323216</td>\n    </tr>\n    <tr>\n      <th>Fedot_Industrial_baseline</th>\n      <td>492.792531</td>\n      <td>492.792531</td>\n      <td>492.792531</td>\n    </tr>\n    <tr>\n      <th>5NN-ED_RMSE</th>\n      <td>424.504623</td>\n      <td>562.413046</td>\n      <td>493.478164</td>\n    </tr>\n    <tr>\n      <th>RandF_RMSE</th>\n      <td>456.359415</td>\n      <td>589.478468</td>\n      <td>527.343198</td>\n    </tr>\n    <tr>\n      <th>RDST_RMSE</th>\n      <td>491.072419</td>\n      <td>575.720140</td>\n      <td>527.927002</td>\n    </tr>\n    <tr>\n      <th>Ridge_RMSE</th>\n      <td>509.119169</td>\n      <td>627.513473</td>\n      <td>574.558434</td>\n    </tr>\n    <tr>\n      <th>ROCKET_RMSE</th>\n      <td>492.914183</td>\n      <td>629.215008</td>\n      <td>575.695827</td>\n    </tr>\n    <tr>\n      <th>FreshPRINCE_RMSE</th>\n      <td>487.642605</td>\n      <td>640.463800</td>\n      <td>582.654848</td>\n    </tr>\n    <tr>\n      <th>DrCIF_RMSE</th>\n      <td>494.923063</td>\n      <td>646.879703</td>\n      <td>587.404402</td>\n    </tr>\n    <tr>\n      <th>SingleInception_RMSE</th>\n      <td>499.449452</td>\n      <td>652.585529</td>\n      <td>588.659847</td>\n    </tr>\n    <tr>\n      <th>ResNet_RMSE</th>\n      <td>489.719817</td>\n      <td>661.954851</td>\n      <td>589.427954</td>\n    </tr>\n    <tr>\n      <th>InceptionT_RMSE</th>\n      <td>506.696071</td>\n      <td>645.057362</td>\n      <td>589.468140</td>\n    </tr>\n    <tr>\n      <th>MultiROCKET_RMSE</th>\n      <td>506.770267</td>\n      <td>663.968054</td>\n      <td>596.389600</td>\n    </tr>\n    <tr>\n      <th>RIST_RMSE</th>\n      <td>560.257921</td>\n      <td>639.787936</td>\n      <td>601.625863</td>\n    </tr>\n    <tr>\n      <th>TSF_RMSE</th>\n      <td>510.156025</td>\n      <td>663.258974</td>\n      <td>604.193543</td>\n    </tr>\n    <tr>\n      <th>RotF_RMSE</th>\n      <td>519.799244</td>\n      <td>670.442108</td>\n      <td>606.662679</td>\n    </tr>\n    <tr>\n      <th>FCN_RMSE</th>\n      <td>499.100408</td>\n      <td>693.068654</td>\n      <td>609.970047</td>\n    </tr>\n    <tr>\n      <th>XGBoost_RMSE</th>\n      <td>526.421564</td>\n      <td>693.574841</td>\n      <td>618.172663</td>\n    </tr>\n    <tr>\n      <th>1NN-DTW_RMSE</th>\n      <td>557.562623</td>\n      <td>755.854409</td>\n      <td>666.809599</td>\n    </tr>\n    <tr>\n      <th>1NN-ED_RMSE</th>\n      <td>529.096971</td>\n      <td>812.146002</td>\n      <td>684.484545</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sort_values('average')"
   ],
   "metadata": {
    "collapsed": false
   }
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [],
   "source": [],
   "metadata": {
    "collapsed": false
   }
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}