notebook/AdversarialValidation.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## References\n",
    "- https://www.kaggle.com/tunguz/ms-malware-adversarial-validation\n",
    "- https://github.com/zygmuntz/adversarial-validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:18:31.939969Z",
     "start_time": "2019-01-17T14:18:30.173504Z"
    },
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "%matplotlib inline\n",
    "import lightgbm as lgb\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn import model_selection, preprocessing, metrics\n",
    "from sklearn.model_selection import StratifiedKFold\n",
    "from sklearn.metrics import roc_auc_score, accuracy_score\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "import os\n",
    "os.environ['KMP_DUPLICATE_LIB_OK']='True' #OSX\n",
    "import gc\n",
    "gc.enable()\n",
    "from pathlib import Path\n",
    "ROOT_PATH = Path(\".\").absolute().parents[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:18:32.820402Z",
     "start_time": "2019-01-17T14:18:32.813989Z"
    },
    "_uuid": "c8e7234e1d85abdf3c257187e2cce3d3800ec610"
   },
   "outputs": [],
   "source": [
    "columns_to_use = ['MachineIdentifier', 'ProductName', 'EngineVersion', 'AppVersion', 'AvSigVersion', 'IsBeta',\n",
    "       'RtpStateBitfield', 'IsSxsPassiveMode', 'DefaultBrowsersIdentifier',\n",
    "       'AVProductStatesIdentifier', 'AVProductsInstalled', 'AVProductsEnabled',\n",
    "       'HasTpm', 'CountryIdentifier', 'CityIdentifier',\n",
    "       'OrganizationIdentifier', 'GeoNameIdentifier',\n",
    "       'LocaleEnglishNameIdentifier', 'Platform', 'Processor', 'OsVer',\n",
    "       'OsBuild', 'OsSuite', 'OsPlatformSubRelease', 'OsBuildLab',\n",
    "       'SkuEdition', 'IsProtected', 'AutoSampleOptIn', 'SMode',\n",
    "       'IeVerIdentifier', 'SmartScreen', 'Firewall', 'UacLuaenable',\n",
    "       'Census_MDC2FormFactor', 'Census_DeviceFamily',\n",
    "       'Census_OEMNameIdentifier', 'Census_OEMModelIdentifier',\n",
    "       'Census_ProcessorCoreCount', 'Census_ProcessorManufacturerIdentifier',\n",
    "       'Census_ProcessorModelIdentifier', 'Census_ProcessorClass',\n",
    "       'Census_PrimaryDiskTotalCapacity', 'Census_PrimaryDiskTypeName',\n",
    "       'Census_SystemVolumeTotalCapacity', 'Census_HasOpticalDiskDrive',\n",
    "       'Census_TotalPhysicalRAM', 'Census_ChassisTypeName',\n",
    "       'Census_InternalPrimaryDiagonalDisplaySizeInInches',\n",
    "       'Census_InternalPrimaryDisplayResolutionHorizontal',\n",
    "       'Census_InternalPrimaryDisplayResolutionVertical',\n",
    "       'Census_PowerPlatformRoleName', 'Census_InternalBatteryType',\n",
    "       'Census_InternalBatteryNumberOfCharges', 'Census_OSVersion',\n",
    "       'Census_OSArchitecture', 'Census_OSBranch', 'Census_OSBuildNumber',\n",
    "       'Census_OSBuildRevision', 'Census_OSEdition', 'Census_OSSkuName',\n",
    "       'Census_OSInstallTypeName', 'Census_OSInstallLanguageIdentifier',\n",
    "       'Census_OSUILocaleIdentifier', 'Census_OSWUAutoUpdateOptionsName',\n",
    "       'Census_IsPortableOperatingSystem', 'Census_GenuineStateName',\n",
    "       'Census_ActivationChannel', 'Census_IsFlightingInternal',\n",
    "       'Census_IsFlightsDisabled', 'Census_FlightRing',\n",
    "       'Census_ThresholdOptIn', 'Census_FirmwareManufacturerIdentifier',\n",
    "       'Census_FirmwareVersionIdentifier', 'Census_IsSecureBootEnabled',\n",
    "       'Census_IsWIMBootEnabled', 'Census_IsVirtualDevice',\n",
    "       'Census_IsTouchEnabled', 'Census_IsPenCapable',\n",
    "       'Census_IsAlwaysOnAlwaysConnectedCapable', 'Wdft_IsGamer',\n",
    "       'Wdft_RegionIdentifier']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:20:52.725572Z",
     "start_time": "2019-01-17T14:18:33.411177Z"
    },
    "_uuid": "0e7763f3340dd66a4e263a80f972965ca2d706a5"
   },
   "outputs": [],
   "source": [
    "new_train_df = pd.read_csv(ROOT_PATH / \"data\" / \"validation\" / \"new_train.csv\", index_col=0, usecols = columns_to_use)\n",
    "new_test_df = pd.read_csv(ROOT_PATH / \"data\" / \"validation\" /  \"new_test.csv\",index_col=0, usecols = columns_to_use)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:20:52.750191Z",
     "start_time": "2019-01-17T14:20:52.729069Z"
    },
    "_uuid": "b2f9b8202537f387d6fb6cedf574354bca3749d3"
   },
   "outputs": [],
   "source": [
    "new_train_df['IsTest'] = 0\n",
    "new_test_df['IsTest'] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:21:18.382407Z",
     "start_time": "2019-01-17T14:20:52.752090Z"
    },
    "_uuid": "c48436bd49eb41836de341f7302caa1b509d0202"
   },
   "outputs": [],
   "source": [
    "adv_df = pd.concat([new_train_df, new_test_df], axis =0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:21:48.203170Z",
     "start_time": "2019-01-17T14:21:18.385607Z"
    }
   },
   "outputs": [],
   "source": [
    "x = adv_df.drop([\"IsTest\"], axis=1)\n",
    "y = adv_df.IsTest\n",
    "param = {'num_leaves': 50,\n",
    "         'min_data_in_leaf': 30, \n",
    "         'objective':'binary',\n",
    "         'max_depth': 4,\n",
    "         'learning_rate': 0.01,\n",
    "         \"min_child_samples\": 20,\n",
    "         \"boosting\": \"gbdt\",\n",
    "         \"feature_fraction\": 0.9,\n",
    "         \"bagging_freq\": 1,\n",
    "         \"bagging_fraction\": 0.9 ,\n",
    "         \"bagging_seed\": 17,\n",
    "         \"metric\": 'auc',\n",
    "         \"lambda_l1\": 0.1,\n",
    "         \"verbosity\": -1,\n",
    "         \"seed\": 42}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:21:48.230115Z",
     "start_time": "2019-01-17T14:21:48.205676Z"
    }
   },
   "outputs": [],
   "source": [
    "predictions = np.zeros(y.shape)\n",
    "folds = StratifiedKFold(n_splits=5, shuffle=True, random_state=5678)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:47:59.566927Z",
     "start_time": "2019-01-17T14:21:48.231759Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "fold 0\n",
      "Training until validation scores don't improve for 25 rounds.\n",
      "[10]\ttraining's auc: 0.965609\tvalid_1's auc: 0.96535\n",
      "[20]\ttraining's auc: 0.969277\tvalid_1's auc: 0.969065\n",
      "[30]\ttraining's auc: 0.97037\tvalid_1's auc: 0.970187\n",
      "[40]\ttraining's auc: 0.970476\tvalid_1's auc: 0.970294\n",
      "[50]\ttraining's auc: 0.970538\tvalid_1's auc: 0.970356\n",
      "[60]\ttraining's auc: 0.970538\tvalid_1's auc: 0.970358\n",
      "[70]\ttraining's auc: 0.970571\tvalid_1's auc: 0.97039\n",
      "[80]\ttraining's auc: 0.97037\tvalid_1's auc: 0.970178\n",
      "[90]\ttraining's auc: 0.970392\tvalid_1's auc: 0.9702\n",
      "[100]\ttraining's auc: 0.970406\tvalid_1's auc: 0.970215\n",
      "Early stopping, best iteration is:\n",
      "[77]\ttraining's auc: 0.970588\tvalid_1's auc: 0.970408\n",
      "AUC: 0.9704079921038696\n",
      "\n",
      "fold 1\n",
      "Training until validation scores don't improve for 25 rounds.\n",
      "[10]\ttraining's auc: 0.965568\tvalid_1's auc: 0.965515\n",
      "[20]\ttraining's auc: 0.969753\tvalid_1's auc: 0.96972\n",
      "[30]\ttraining's auc: 0.970332\tvalid_1's auc: 0.970303\n",
      "[40]\ttraining's auc: 0.970459\tvalid_1's auc: 0.970429\n",
      "[50]\ttraining's auc: 0.970506\tvalid_1's auc: 0.970475\n",
      "[60]\ttraining's auc: 0.970511\tvalid_1's auc: 0.97048\n",
      "[70]\ttraining's auc: 0.970521\tvalid_1's auc: 0.97049\n",
      "[80]\ttraining's auc: 0.97034\tvalid_1's auc: 0.970303\n",
      "[90]\ttraining's auc: 0.970368\tvalid_1's auc: 0.970327\n",
      "[100]\ttraining's auc: 0.970368\tvalid_1's auc: 0.970327\n",
      "Early stopping, best iteration is:\n",
      "[76]\ttraining's auc: 0.970565\tvalid_1's auc: 0.970534\n",
      "AUC: 0.9705337412777099\n",
      "\n",
      "fold 2\n",
      "Training until validation scores don't improve for 25 rounds.\n",
      "[10]\ttraining's auc: 0.963493\tvalid_1's auc: 0.963411\n",
      "[20]\ttraining's auc: 0.964602\tvalid_1's auc: 0.964488\n",
      "[30]\ttraining's auc: 0.96909\tvalid_1's auc: 0.968971\n",
      "[40]\ttraining's auc: 0.969356\tvalid_1's auc: 0.969239\n",
      "[50]\ttraining's auc: 0.96947\tvalid_1's auc: 0.969353\n",
      "[60]\ttraining's auc: 0.969636\tvalid_1's auc: 0.969518\n",
      "[70]\ttraining's auc: 0.969761\tvalid_1's auc: 0.969643\n",
      "[80]\ttraining's auc: 0.969795\tvalid_1's auc: 0.969678\n",
      "[90]\ttraining's auc: 0.970579\tvalid_1's auc: 0.970506\n",
      "[100]\ttraining's auc: 0.970577\tvalid_1's auc: 0.970504\n",
      "[110]\ttraining's auc: 0.970591\tvalid_1's auc: 0.970518\n",
      "Early stopping, best iteration is:\n",
      "[86]\ttraining's auc: 0.970808\tvalid_1's auc: 0.970722\n",
      "AUC: 0.970722497594744\n",
      "\n",
      "fold 3\n",
      "Training until validation scores don't improve for 25 rounds.\n",
      "[10]\ttraining's auc: 0.963414\tvalid_1's auc: 0.963726\n",
      "[20]\ttraining's auc: 0.964522\tvalid_1's auc: 0.96481\n",
      "[30]\ttraining's auc: 0.968758\tvalid_1's auc: 0.969027\n",
      "[40]\ttraining's auc: 0.969049\tvalid_1's auc: 0.969312\n",
      "[50]\ttraining's auc: 0.969206\tvalid_1's auc: 0.969462\n",
      "[60]\ttraining's auc: 0.969403\tvalid_1's auc: 0.969652\n",
      "[70]\ttraining's auc: 0.969459\tvalid_1's auc: 0.969705\n",
      "[80]\ttraining's auc: 0.969494\tvalid_1's auc: 0.969739\n",
      "[90]\ttraining's auc: 0.970373\tvalid_1's auc: 0.970598\n",
      "[100]\ttraining's auc: 0.970368\tvalid_1's auc: 0.970595\n",
      "[110]\ttraining's auc: 0.970606\tvalid_1's auc: 0.97082\n",
      "[120]\ttraining's auc: 0.970744\tvalid_1's auc: 0.970958\n",
      "[130]\ttraining's auc: 0.972312\tvalid_1's auc: 0.972509\n",
      "[140]\ttraining's auc: 0.972295\tvalid_1's auc: 0.972504\n",
      "[150]\ttraining's auc: 0.972743\tvalid_1's auc: 0.97294\n",
      "[160]\ttraining's auc: 0.972568\tvalid_1's auc: 0.972769\n",
      "[170]\ttraining's auc: 0.972579\tvalid_1's auc: 0.972781\n",
      "Early stopping, best iteration is:\n",
      "[151]\ttraining's auc: 0.972743\tvalid_1's auc: 0.972941\n",
      "AUC: 0.9729406974768587\n",
      "\n",
      "fold 4\n",
      "Training until validation scores don't improve for 25 rounds.\n",
      "[10]\ttraining's auc: 0.963463\tvalid_1's auc: 0.963529\n",
      "[20]\ttraining's auc: 0.964559\tvalid_1's auc: 0.964661\n",
      "[30]\ttraining's auc: 0.968803\tvalid_1's auc: 0.968864\n",
      "[40]\ttraining's auc: 0.969093\tvalid_1's auc: 0.969153\n",
      "[50]\ttraining's auc: 0.969215\tvalid_1's auc: 0.969276\n",
      "[60]\ttraining's auc: 0.969456\tvalid_1's auc: 0.969522\n",
      "[70]\ttraining's auc: 0.969501\tvalid_1's auc: 0.969568\n",
      "[80]\ttraining's auc: 0.969533\tvalid_1's auc: 0.9696\n",
      "[90]\ttraining's auc: 0.970419\tvalid_1's auc: 0.970491\n",
      "[100]\ttraining's auc: 0.970415\tvalid_1's auc: 0.970488\n",
      "[110]\ttraining's auc: 0.970433\tvalid_1's auc: 0.970506\n",
      "Early stopping, best iteration is:\n",
      "[86]\ttraining's auc: 0.97064\tvalid_1's auc: 0.9707\n",
      "AUC: 0.9706998213004912\n",
      "\n"
     ]
    }
   ],
   "source": [
    "for fold_, (train_idx, valid_idx) in enumerate(folds.split(x.values, y.values)):\n",
    "    print(\"fold {}\".format(fold_))\n",
    "    train_data = lgb.Dataset(x.iloc[train_idx], label=y.iloc[train_idx])\n",
    "    valid_data = lgb.Dataset(x.iloc[valid_idx], label=y.iloc[valid_idx])\n",
    "    num_round = 1000\n",
    "    clf = lgb.train(param, train_data, num_round, valid_sets=[train_data, valid_data], verbose_eval=10, early_stopping_rounds=25)\n",
    "    clf.save_model(\"../data/model/lgbm_for_adv_validation_{}.model\".format(fold_))\n",
    "    predict = clf.predict(x.iloc[valid_idx], num_iteration=clf.best_iteration)\n",
    "    auc = roc_auc_score(y.iloc[valid_idx], predict)\n",
    "    print(\"AUC: {}\\n\".format(auc))\n",
    "    predictions[valid_idx] = predict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:49:36.032175Z",
     "start_time": "2019-01-17T14:48:31.757470Z"
    }
   },
   "outputs": [],
   "source": [
    "adv_df[\"Predict\"] = predictions\n",
    "i = predictions.argsort()\n",
    "train_sorted = adv_df[\"Predict\"].iloc[i]\n",
    "train_sorted = pd.merge(pd.DataFrame(train_sorted), pd.DataFrame(adv_df[\"IsTest\"]), on=\"MachineIdentifier\")\n",
    "train_sorted = train_sorted.loc[train_sorted.IsTest == 0]\n",
    "train_sorted = train_sorted.drop(\"IsTest\", axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:49:36.745683Z",
     "start_time": "2019-01-17T14:49:36.034893Z"
    }
   },
   "outputs": [],
   "source": [
    "del new_train_df, new_test_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:51:41.757503Z",
     "start_time": "2019-01-17T14:51:12.576592Z"
    }
   },
   "outputs": [],
   "source": [
    "detection_df = pd.read_csv(ROOT_PATH / \"data\" / \"validation\" / \"new_train.csv\",\n",
    "                           index_col=0,\n",
    "                           usecols = [\"MachineIdentifier\", \"HasDetections\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:53:57.548101Z",
     "start_time": "2019-01-17T14:53:40.928508Z"
    }
   },
   "outputs": [],
   "source": [
    "joined_df = pd.merge(train_sorted, detection_df, on=\"MachineIdentifier\")\n",
    "joined_df = joined_df.reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:53:58.362286Z",
     "start_time": "2019-01-17T14:53:57.551149Z"
    }
   },
   "outputs": [],
   "source": [
    "detected_df = joined_df.query(\"HasDetections==1\")\n",
    "not_detected_df = joined_df.query(\"HasDetections==0\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:54:14.404592Z",
     "start_time": "2019-01-17T14:54:00.938497Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(detected_df[\"Predict\"])\n",
    "plt.title(\"Detected Prediction\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T14:54:27.609019Z",
     "start_time": "2019-01-17T14:54:14.406649Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(not_detected_df[\"Predict\"])\n",
    "plt.title(\"Not Detected Prediction\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T15:01:53.108417Z",
     "start_time": "2019-01-17T15:01:52.339673Z"
    }
   },
   "outputs": [],
   "source": [
    "detected_predict_df = detected_df[[\"MachineIdentifier\", \"Predict\"]]\n",
    "not_detected_predict_df = not_detected_df[[\"MachineIdentifier\", \"Predict\"]]\n",
    "joined_df[\"Validation\"] = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-17T15:02:46.693920Z",
     "start_time": "2019-01-17T15:02:06.407585Z"
    }
   },
   "outputs": [],
   "source": [
    "for i in range(5):\n",
    "    detected_fold = detected_predict_df.iloc[i::5]\n",
    "    not_detected_fold = not_detected_predict_df[i::5]\n",
    "    joined_df.loc[joined_df[\"MachineIdentifier\"].isin(detected_fold[\"MachineIdentifier\"]), \"Validation\"] = i\n",
    "    joined_df.loc[joined_df[\"MachineIdentifier\"].isin(not_detected_fold[\"MachineIdentifier\"]), \"Validation\"] = i"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-12T12:39:58.631400Z",
     "start_time": "2019-01-12T12:39:58.615045Z"
    },
    "deletable": false,
    "editable": false,
    "run_control": {
     "frozen": true
    }
   },
   "outputs": [],
   "source": [
    "del joined_df[\"Predict\"], joined_df[\"HasDetections\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-12T12:40:50.933088Z",
     "start_time": "2019-01-12T12:40:24.718284Z"
    },
    "deletable": false,
    "editable": false,
    "run_control": {
     "frozen": true
    }
   },
   "outputs": [],
   "source": [
    "joined_df.to_csv(ROOT_PATH / \"data\" / \"validations\" / \"adv_train.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "start_time": "2019-01-17T15:03:43.820Z"
    }
   },
   "outputs": [],
   "source": [
    "joined_df.to_csv(ROOT_PATH / \"data\" / \"validation\" / \"adv_train.csv\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.7"
  },
  "notify_time": "10",
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}