Bagging.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Bagging and Random Forest Models\n",
    "\n",
    "Using **ensemble methods** can greatly improve the results achieved with weak machine learning algorithms, also called **weak learners**. Ensemble methods achieve better performance by aggregating the results of many statistically independent models. This process averages out the errors and produces a final better prediction. \n",
    "\n",
    "In this lab you will work with a widely used ensemble method known as **bootstrap aggregating** or simply **bagging**. Bagging follows a simple procedure:\n",
    "1. N learners (machine learning models) are defined. \n",
    "2. N subsamples of the training data are created by **Bernoulli sampling with replacement**.\n",
    "3. The N learners are trained on the subsamples of the training data.\n",
    "4. The ensemble is scored by averaging, or taking a majority vote, of the predictions from the N learners.\n",
    "\n",
    "**Classification and regression tree models** are most typically used with bagging methods. The most common such algorithm is know as the **random forest**. The random forest method is highly scalable and generally produces good results, even for complex problems. \n",
    "\n",
    "Classification and regression trees tend to be robust to noise or outliers in the training data. This is true for the random forest algorithm as well. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As a first step, execute the code in the cell below to load the required packages to run the rest of this notebook. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn import preprocessing\n",
    "from sklearn import datasets ## Get dataset from sklearn\n",
    "import sklearn.model_selection as ms\n",
    "import sklearn.metrics as sklm\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import numpy.random as nr\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To get a feel for these data, you will now load and plot them. The code in the cell below does the following:\n",
    "\n",
    "1. Loads the iris data as a Pandas data frame. \n",
    "2. Adds column names to the data frame.\n",
    "3. Displays all 4 possible scatter plot views of the data. \n",
    "\n",
    "Execute this code and examine the results. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x864 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_iris(iris):\n",
    "    '''Function to plot iris data by type'''\n",
    "    setosa = iris[iris['Species'] == 'setosa']\n",
    "    versicolor = iris[iris['Species'] == 'versicolor']\n",
    "    virginica = iris[iris['Species'] == 'virginica']\n",
    "    fig, ax = plt.subplots(2, 2, figsize=(12,12))\n",
    "    x_ax = ['Sepal_Length', 'Sepal_Width']\n",
    "    y_ax = ['Petal_Length', 'Petal_Width']\n",
    "    for i in range(2):\n",
    "        for j in range(2):\n",
    "            ax[i,j].scatter(setosa[x_ax[i]], setosa[y_ax[j]], marker = 'x')\n",
    "            ax[i,j].scatter(versicolor[x_ax[i]], versicolor[y_ax[j]], marker = 'o')\n",
    "            ax[i,j].scatter(virginica[x_ax[i]], virginica[y_ax[j]], marker = '+')\n",
    "            ax[i,j].set_xlabel(x_ax[i])\n",
    "            ax[i,j].set_ylabel(y_ax[j])\n",
    "            \n",
    "## Import the dataset from sklearn.datasets\n",
    "iris = datasets.load_iris()\n",
    "\n",
    "## Create a data frame from the dictionary\n",
    "species = [iris.target_names[x] for x in iris.target]\n",
    "iris = pd.DataFrame(iris['data'], columns = ['Sepal_Length', 'Sepal_Width', 'Petal_Length', 'Petal_Width'])\n",
    "iris['Species'] = species\n",
    "\n",
    "## Plot views of the iris data            \n",
    "plot_iris(iris) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "You can see that Setosa (in blue) is well separated from the other two categories. The Versicolor (in orange) and the Virginica (in green) show considerable overlap. The question is how well our classifier will separate these categories. \n",
    "\n",
    "Scikit Learn classifiers require numerically coded numpy arrays for the features and as a label. The code in the cell below does the following processing:\n",
    "1. Creates a numpy array of the features.\n",
    "2. Numerically codes the label using a dictionary lookup, and converts it to a numpy array. \n",
    "\n",
    "Execute this code."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "Features = np.array(iris[['Sepal_Length', 'Sepal_Width', 'Petal_Length', 'Petal_Width']])\n",
    "\n",
    "levels = {'setosa':0, 'versicolor':1, 'virginica':2}\n",
    "Labels =  np.array([levels[x] for x in iris['Species']])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, execute the code in the cell below to split the dataset into test and training set. Notice that unusually, 100 of the 150 cases are being used as the test dataset. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Randomly sample cases to create independent training and test data\n",
    "nr.seed(1115)\n",
    "indx = range(Features.shape[0])\n",
    "indx = ms.train_test_split(indx, test_size = 100)\n",
    "X_train = Features[indx[0],:]\n",
    "y_train = np.ravel(Labels[indx[0]])\n",
    "X_test = Features[indx[1],:]\n",
    "y_test = np.ravel(Labels[indx[1]])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As is always the case with machine learning, numeric features  must be scaled. The code in the cell below performs the following processing:\n",
    "\n",
    "1. A Zscore scale object is defined using the `StandarScaler` function from the Scikit Learn preprocessing package. \n",
    "2. The scaler is fit to the training features. Subsequently, this scaler is used to apply the same scaling to the test data and in production. \n",
    "3. The training features are scaled using the `transform` method. \n",
    "\n",
    "Execute this code."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "scale = preprocessing.StandardScaler()\n",
    "scale.fit(X_train)\n",
    "X_train = scale.transform(X_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now you will define and fit a random forest model. The code in the cell below defines random forest model with 5 trees using the `RandomForestClassifer` function from the Scikit Learn ensemble  package, and then fits the model. Execute this code."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier(bootstrap=True, class_weight=None, criterion='gini',\n",
       "                       max_depth=None, max_features='auto', max_leaf_nodes=None,\n",
       "                       min_impurity_decrease=0.0, min_impurity_split=None,\n",
       "                       min_samples_leaf=1, min_samples_split=2,\n",
       "                       min_weight_fraction_leaf=0.0, n_estimators=5,\n",
       "                       n_jobs=None, oob_score=False, random_state=None,\n",
       "                       verbose=0, warm_start=False)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nr.seed(444)\n",
    "rf_clf = RandomForestClassifier(n_estimators=5)\n",
    "rf_clf.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Notice that the many hyperparameters of the random forest model object are displayed. \n",
    "\n",
    "Next, the code in the cell below performs the following processing to score the test data subset:\n",
    "1. The test features are scaled using the scaler computed for the training features. \n",
    "2. The `predict` method is used to compute the scores from the scaled features. \n",
    "\n",
    "Execute this code. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = scale.transform(X_test)\n",
    "scores = rf_clf.predict(X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It is time to evaluate the model results. Keep in mind that the problem has been made difficult deliberately, by having more test cases than training cases. \n",
    "\n",
    "The iris data has three species categories. Therefore it is necessary to use evaluation code for a three category problem. The function in the cell below extends code from pervious labs to deal with a three category problem. Execute this code and examine the results."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                 Confusion matrix\n",
      "                 Score Setosa   Score Versicolor    Score Virginica\n",
      "Actual Setosa          35                0                 0\n",
      "Actual Versicolor       0               30                 4\n",
      "Actual Vriginica        0                2                29\n",
      "\n",
      "Accuracy        0.94\n",
      " \n",
      "          Setosa  Versicolor  Virginica\n",
      "Num case   35.00     34.00      31.00\n",
      "Precision   1.00      0.94       0.88\n",
      "Recall      1.00      0.88       0.94\n",
      "F1          1.00      0.91       0.91\n"
     ]
    }
   ],
   "source": [
    "def print_metrics_3(labels, scores):\n",
    "   \n",
    "    conf = sklm.confusion_matrix(labels, scores)\n",
    "    print('                 Confusion matrix')\n",
    "    print('                 Score Setosa   Score Versicolor    Score Virginica')\n",
    "    print('Actual Setosa      %6d' % conf[0,0] + '            %5d' % conf[0,1] + '             %5d' % conf[0,2])\n",
    "    print('Actual Versicolor  %6d' % conf[1,0] + '            %5d' % conf[1,1] + '             %5d' % conf[1,2])\n",
    "    print('Actual Vriginica   %6d' % conf[2,0] + '            %5d' % conf[2,1] + '             %5d' % conf[2,2])\n",
    "    ## Now compute and display the accuracy and metrics\n",
    "    print('')\n",
    "    print('Accuracy        %0.2f' % sklm.accuracy_score(labels, scores))\n",
    "    metrics = sklm.precision_recall_fscore_support(labels, scores)\n",
    "    print(' ')\n",
    "    print('          Setosa  Versicolor  Virginica')\n",
    "    print('Num case   %0.2f' % metrics[3][0] + '     %0.2f' % metrics[3][1] + '      %0.2f' % metrics[3][2])\n",
    "    print('Precision   %0.2f' % metrics[0][0] + '      %0.2f' % metrics[0][1] + '       %0.2f' % metrics[0][2])\n",
    "    print('Recall      %0.2f' % metrics[1][0] + '      %0.2f' % metrics[1][1] + '       %0.2f' % metrics[1][2])\n",
    "    print('F1          %0.2f' % metrics[2][0] + '      %0.2f' % metrics[2][1] + '       %0.2f' % metrics[2][2])\n",
    "    \n",
    "print_metrics_3(y_test, scores)    "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Examine these results. Notice the following:\n",
    "1. The confusion matrix has dimension 3X3. You can see that most cases are correctly classified with only a few errors. \n",
    "2. The overall accuracy is 0.94. Since the classes are roughly balanced, this metric indicates relatively good performance of the classifier, particularly since it was only trained on 50 cases. \n",
    "3. The precision, recall and  F1 for each of the classes is quite good.\n",
    "|\n",
    "To get a better feel for what the classifier is doing, the code in the cell below displays a set of plots showing correctly (as '+') and incorrectly (as 'o') cases, with the species color-coded. Execute this code and examine the results. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x864 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def plot_iris_score(iris, y_test, scores):\n",
    "    '''Function to plot iris data by type'''\n",
    "    ## Find correctly and incorrectly classified cases\n",
    "    true = np.equal(scores, y_test).astype(int)\n",
    "    \n",
    "    ## Create data frame from the test data\n",
    "    iris = pd.DataFrame(iris)\n",
    "    levels = {0:'setosa', 1:'versicolor', 2:'virginica'}\n",
    "    iris['Species'] = [levels[x] for x in y_test]\n",
    "    iris.columns = ['Sepal_Length', 'Sepal_Width', 'Petal_Length', 'Petal_Width', 'Species']\n",
    "    \n",
    "    ## Set up for the plot\n",
    "    fig, ax = plt.subplots(2, 2, figsize=(12,12))\n",
    "    markers = ['o', '+']\n",
    "    x_ax = ['Sepal_Length', 'Sepal_Width']\n",
    "    y_ax = ['Petal_Length', 'Petal_Width']\n",
    "    \n",
    "    for t in range(2): # loop over correct and incorect classifications\n",
    "        setosa = iris[(iris['Species'] == 'setosa') & (true == t)]\n",
    "        versicolor = iris[(iris['Species'] == 'versicolor') & (true == t)]\n",
    "        virginica = iris[(iris['Species'] == 'virginica') & (true == t)]\n",
    "        # loop over all the dimensions\n",
    "        for i in range(2):\n",
    "            for j in range(2):\n",
    "                ax[i,j].scatter(setosa[x_ax[i]], setosa[y_ax[j]], marker = markers[t], color = 'blue')\n",
    "                ax[i,j].scatter(versicolor[x_ax[i]], versicolor[y_ax[j]], marker = markers[t], color = 'orange')\n",
    "                ax[i,j].scatter(virginica[x_ax[i]], virginica[y_ax[j]], marker = markers[t], color = 'green')\n",
    "                ax[i,j].set_xlabel(x_ax[i])\n",
    "                ax[i,j].set_ylabel(y_ax[j])\n",
    "\n",
    "plot_iris_score(X_test, y_test, scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Examine these plots. You can see how the classifier has divided the feature space between the classes. Notice that most of the errors occur in the overlap region between Virginica and Versicolor. This behavior is to be expected.  \n",
    "\n",
    "Is it possible that a random forest model with more trees would separate these cases better? The code in the cell below uses a model with 40 trees (estimators). This model is fit with the training data and displays the evaluation of the model. \n",
    "\n",
    "Execute this code and answer **Question 1** on the course page."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                 Confusion matrix\n",
      "                 Score Setosa   Score Versicolor    Score Virginica\n",
      "Actual Setosa          35                0                 0\n",
      "Actual Versicolor       0               32                 2\n",
      "Actual Vriginica        0                3                28\n",
      "\n",
      "Accuracy        0.95\n",
      " \n",
      "          Setosa  Versicolor  Virginica\n",
      "Num case   35.00     34.00      31.00\n",
      "Precision   1.00      0.91       0.93\n",
      "Recall      1.00      0.94       0.90\n",
      "F1          1.00      0.93       0.92\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x864 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "nr.seed(444)\n",
    "rf_clf = RandomForestClassifier(n_estimators=40)\n",
    "rf_clf.fit(X_train, y_train)\n",
    "scores = rf_clf.predict(X_test)\n",
    "print_metrics_3(y_test, scores) \n",
    "plot_iris_score(X_test, y_test, scores)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These results are slightly better than for the model with 5 trees. However, this small difference in performance is unlikely to be significant. \n",
    "\n",
    "Like most tree-based models, random forest models have a nice property that **feature importance** is computed during model training. Feature importance can be used as a feature selection method. \n",
    "\n",
    "Execute the code in the cell below to display a plot of the feature importance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0.04644949 0.10044804 0.35086919 0.50223328]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Feature importance')"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "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": [
    "importance = rf_clf.feature_importances_\n",
    "print(importance)\n",
    "plt.bar(range(4), importance, tick_label = ['Sepal_Length', 'Sepal_Width', 'Petal_Length', 'Petal_Width'])\n",
    "plt.xticks(rotation=90)\n",
    "plt.ylabel('Feature importance')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Examine the plot displayed above. Notice that the Speal_Lenght and Sepal_Width have rather low importance. \n",
    "\n",
    "Should these features be dropped from the model? To find out, you will create a model with a reduced feature set and compare the results. As a first step, execute the code in the cell below to create training and test datasets using the reduced features."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "## Create reduced feature set\n",
    "Features = np.array(iris[['Petal_Length', 'Petal_Width']])\n",
    "\n",
    "## Randomly sample cases to create independent training and test data\n",
    "nr.seed(1115)\n",
    "indx = range(Features.shape[0])\n",
    "indx = ms.train_test_split(indx, test_size = 100)\n",
    "X_train = Features[indx[0],:]\n",
    "y_train = np.ravel(Labels[indx[0]])\n",
    "X_test = Features[indx[1],:]\n",
    "y_test = np.ravel(Labels[indx[1]])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, execute the code in the cell below to define the model, fit the model, score the model and print the results. \n",
    "\n",
    "Once you have executed the code, answer **Question 2** on the course page."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                 Confusion matrix\n",
      "                 Score Setosa   Score Versicolor    Score Virginica\n",
      "Actual Setosa          35                0                 0\n",
      "Actual Versicolor       0               30                 4\n",
      "Actual Vriginica        0                1                30\n",
      "\n",
      "Accuracy        0.95\n",
      " \n",
      "          Setosa  Versicolor  Virginica\n",
      "Num case   35.00     34.00      31.00\n",
      "Precision   1.00      0.97       0.88\n",
      "Recall      1.00      0.88       0.97\n",
      "F1          1.00      0.92       0.92\n"
     ]
    }
   ],
   "source": [
    "nr.seed(444)\n",
    "rf_clf = RandomForestClassifier(n_estimators=40)\n",
    "rf_clf.fit(X_train, y_train)\n",
    "scores = rf_clf.predict(X_test)\n",
    "print_metrics_3(y_test, scores) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "These results are nearly the same as those obtained with the full feature set. In all likelihood, there is no significant difference. Given that a simpler model is more likely to generalize, this model is preferred. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Summary\n",
    "\n",
    "In this lab you have accomplished the following:\n",
    "1. Used a random forest model to classify the cases of the iris data. A model with more trees had marginally lower error rates, but likely not significantly different.\n",
    "2. Applied feature importance was used for feature selection with the iris data. The model created and evaluated with the reduced feature set had essentially the same performance as the model with more features.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}