03-classification-multiclass.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# Multiclass Classification\n",
        "\n",
        "In the last notebook, we looked at binary classification. This  works well when the data observations belong to one of two classes or categories, such as \"True\" or \"False\". When the data can be categorized into more than two classes, you must use a multiclass classification algorithm.\n",
        "\n",
        "Multiclass classification can be thought of as a combination of multiple binary classifiers. There are two ways in which you approach the problem:\n",
        "\n",
        "- **One vs Rest (OVR)**, in which a classifier is created for each possible class value, with a positive outcome for cases where the prediction is *this* class, and negative predictions for cases where the prediction is any other class. A classification problem with four possible shape classes (*square*, *circle*, *triangle*, *hexagon*) would require four classifiers that predict:\n",
        "    - *square* or not\n",
        "    - *circle* or not\n",
        "    - *triangle* or not\n",
        "    - *hexagon* or not\n",
        "    \n",
        "- **One vs One (OVO)**, in which a classifier for each possible pair of classes is created. The classification problem with four shape classes would require the following binary classifiers:\n",
        "    - *square* or *circle*\n",
        "    - *square* or *triangle*\n",
        "    - *square* or *hexagon*\n",
        "    - *circle* or *triangle*\n",
        "    - *circle* or *hexagon*\n",
        "    - *triangle* or *hexagon*\n",
        "\n",
        "In both approaches, the overall model that combines the classifiers generates a vector of predictions in which the probabilities generated from the individual binary classifiers are used to determine which class to predict.\n",
        "\n",
        "Fortunately, in most machine learning frameworks, including scikit-learn, implementing a multiclass classification model is not significantly more complex than binary classification - and in most cases, the estimators used for binary classification implicitly support multiclass classification by abstracting an OVR algorithm, an OVO algorithm, or by allowing a choice of either.\n",
        "\n",
        "> **More Information**: To learn more about estimator support for multiclass classification in Scikit-Learn, see the [Scikit-Learn documentation](https://scikit-learn.org/stable/modules/multiclass.html).\n",
        "\n",
        "### Explore the data\n",
        "\n",
        "Let's start by examining a dataset that contains observations of multiple classes. We'll use a dataset that contains observations of three different species of penguin.\n",
        "\n",
        "> **Citation**: The penguins dataset used in the this exercise is a subset of data collected and made available by [Dr. Kristen\n",
        "Gorman](https://www.uaf.edu/cfos/people/faculty/detail/kristen-gorman.php)\n",
        "and the [Palmer Station, Antarctica LTER](https://pal.lternet.edu/), a\n",
        "member of the [Long Term Ecological Research\n",
        "Network](https://lternet.edu/)."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "\n",
        "# load the training dataset\n",
        "!wget https://raw.githubusercontent.com/MicrosoftDocs/mslearn-introduction-to-machine-learning/main/Data/ml-basics/penguins.csv\n",
        "penguins = pd.read_csv('penguins.csv')\n",
        "\n",
        "# Display a random sample of 10 observations\n",
        "sample = penguins.sample(10)\n",
        "sample"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "--2023-08-13 20:48:31--  https://raw.githubusercontent.com/MicrosoftDocs/mslearn-introduction-to-machine-learning/main/Data/ml-basics/penguins.csv\r\nResolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.111.133, 185.199.110.133, 185.199.109.133, ...\r\nConnecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.111.133|:443... connected.\r\nHTTP request sent, awaiting response... 200 OK\r\nLength: 7086 (6.9K) [text/plain]\r\nSaving to: ‘penguins.csv’\r\n\r\n\rpenguins.csv          0%[                    ]       0  --.-KB/s               \rpenguins.csv        100%[===================>]   6.92K  --.-KB/s    in 0s      \r\n\r\n2023-08-13 20:48:31 (38.7 MB/s) - ‘penguins.csv’ saved [7086/7086]\r\n\r\n"
        },
        {
          "output_type": "execute_result",
          "execution_count": 1,
          "data": {
            "text/plain": "     CulmenLength  CulmenDepth  FlipperLength  BodyMass  Species\n194          45.3         13.7          210.0    4300.0        1\n5            39.3         20.6          190.0    3650.0        0\n73           45.8         18.9          197.0    4150.0        0\n40           36.5         18.0          182.0    3150.0        0\n129          44.1         18.0          210.0    4000.0        0\n110          38.1         16.5          198.0    3825.0        0\n94           36.2         17.3          187.0    3300.0        0\n273          50.4         15.7          222.0    5750.0        1\n42           36.0         18.5          186.0    3100.0        0\n193          49.6         16.0          225.0    5700.0        1",
            "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>CulmenLength</th>\n      <th>CulmenDepth</th>\n      <th>FlipperLength</th>\n      <th>BodyMass</th>\n      <th>Species</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>194</th>\n      <td>45.3</td>\n      <td>13.7</td>\n      <td>210.0</td>\n      <td>4300.0</td>\n      <td>1</td>\n    </tr>\n    <tr>\n      <th>5</th>\n      <td>39.3</td>\n      <td>20.6</td>\n      <td>190.0</td>\n      <td>3650.0</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>73</th>\n      <td>45.8</td>\n      <td>18.9</td>\n      <td>197.0</td>\n      <td>4150.0</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>40</th>\n      <td>36.5</td>\n      <td>18.0</td>\n      <td>182.0</td>\n      <td>3150.0</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>129</th>\n      <td>44.1</td>\n      <td>18.0</td>\n      <td>210.0</td>\n      <td>4000.0</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>110</th>\n      <td>38.1</td>\n      <td>16.5</td>\n      <td>198.0</td>\n      <td>3825.0</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>94</th>\n      <td>36.2</td>\n      <td>17.3</td>\n      <td>187.0</td>\n      <td>3300.0</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>273</th>\n      <td>50.4</td>\n      <td>15.7</td>\n      <td>222.0</td>\n      <td>5750.0</td>\n      <td>1</td>\n    </tr>\n    <tr>\n      <th>42</th>\n      <td>36.0</td>\n      <td>18.5</td>\n      <td>186.0</td>\n      <td>3100.0</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>193</th>\n      <td>49.6</td>\n      <td>16.0</td>\n      <td>225.0</td>\n      <td>5700.0</td>\n      <td>1</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
          },
          "metadata": {}
        }
      ],
      "execution_count": 1,
      "metadata": {}
    },
    {
      "cell_type": "markdown",
      "source": [
        "The dataset contains the following columns:\n",
        "* **CulmenLength**: The length in mm of the penguin's culmen (bill).\n",
        "* **CulmenDepth**: The depth in mm of the penguin's culmen.\n",
        "* **FlipperLength**: The length in mm of the penguin's flipper.\n",
        "* **BodyMass**: The body mass of the penguin in grams.\n",
        "* **Species**: An integer value that represents the species of the penguin.\n",
        "\n",
        "The **Species** column is the label we want to train a model to predict. The dataset includes three possible species, which are encoded as 0, 1, and 2. The actual species names are revealed by the code below:"
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "penguin_classes = ['Adelie', 'Gentoo', 'Chinstrap']\n",
        "print(sample.columns[0:5].values, 'SpeciesName')\n",
        "for index, row in penguins.sample(10).iterrows():\n",
        "    print('[',row[0], row[1], row[2], row[3], int(row[4]),']',penguin_classes[int(row[4])])"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "['CulmenLength' 'CulmenDepth' 'FlipperLength' 'BodyMass' 'Species'] SpeciesName\n[ 45.8 14.2 219.0 4700.0 1 ] Gentoo\n[ 42.6 13.7 213.0 4950.0 1 ] Gentoo\n[ 50.0 19.5 196.0 3900.0 2 ] Chinstrap\n[ 42.4 17.3 181.0 3600.0 2 ] Chinstrap\n[ 42.8 18.5 195.0 4250.0 0 ] Adelie\n[ 46.4 15.6 221.0 5000.0 1 ] Gentoo\n[ 41.1 17.6 182.0 3200.0 0 ] Adelie\n[ 52.2 18.8 197.0 3450.0 2 ] Chinstrap\n[ 50.1 15.0 225.0 5000.0 1 ] Gentoo\n[ 41.1 18.6 189.0 3325.0 0 ] Adelie\n"
        }
      ],
      "execution_count": 2,
      "metadata": {
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now that we know what the features and labels in the data represent, let's explore the dataset. First, let's see if there are any missing (*null*) values."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "# Count the number of null values for each column\n",
        "penguins.isnull().sum()"
      ],
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 3,
          "data": {
            "text/plain": "CulmenLength     2\nCulmenDepth      2\nFlipperLength    2\nBodyMass         2\nSpecies          0\ndtype: int64"
          },
          "metadata": {}
        }
      ],
      "execution_count": 3,
      "metadata": {
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "It looks like there are some missing feature values, but no missing labels. Let's dig a little deeper and see the rows that contain nulls."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "# Show rows containing nulls\n",
        "penguins[penguins.isnull().any(axis=1)]"
      ],
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 4,
          "data": {
            "text/plain": "     CulmenLength  CulmenDepth  FlipperLength  BodyMass  Species\n3             NaN          NaN            NaN       NaN        0\n271           NaN          NaN            NaN       NaN        1",
            "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>CulmenLength</th>\n      <th>CulmenDepth</th>\n      <th>FlipperLength</th>\n      <th>BodyMass</th>\n      <th>Species</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>3</th>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>271</th>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>NaN</td>\n      <td>1</td>\n    </tr>\n  </tbody>\n</table>\n</div>"
          },
          "metadata": {}
        }
      ],
      "execution_count": 4,
      "metadata": {
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "There are two rows that contain no feature values at all (*NaN* stands for \"not a number\"), so these won't be useful in training a model. Let's discard them from the dataset."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "# Drop rows containing NaN values\n",
        "penguins=penguins.dropna()\n",
        "#Confirm there are now no nulls\n",
        "penguins.isnull().sum()\n"
      ],
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 5,
          "data": {
            "text/plain": "CulmenLength     0\nCulmenDepth      0\nFlipperLength    0\nBodyMass         0\nSpecies          0\ndtype: int64"
          },
          "metadata": {}
        }
      ],
      "execution_count": 5,
      "metadata": {}
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now that we've dealt with the missing values, let's explore how the features relate to the label by creating some box charts."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from matplotlib import pyplot as plt\n",
        "%matplotlib inline\n",
        "\n",
        "penguin_features = ['CulmenLength','CulmenDepth','FlipperLength','BodyMass']\n",
        "penguin_label = 'Species'\n",
        "for col in penguin_features:\n",
        "    penguins.boxplot(column=col, by=penguin_label, figsize=(6,6))\n",
        "    plt.title(col)\n",
        "plt.show()"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": "/anaconda/envs/azureml_py38/lib/python3.8/site-packages/matplotlib/cbook/__init__.py:1376: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n  X = np.atleast_1d(X.T if isinstance(X, np.ndarray) else np.asarray(X))\n/anaconda/envs/azureml_py38/lib/python3.8/site-packages/matplotlib/cbook/__init__.py:1376: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n  X = np.atleast_1d(X.T if isinstance(X, np.ndarray) else np.asarray(X))\n/anaconda/envs/azureml_py38/lib/python3.8/site-packages/matplotlib/cbook/__init__.py:1376: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n  X = np.atleast_1d(X.T if isinstance(X, np.ndarray) else np.asarray(X))\n/anaconda/envs/azureml_py38/lib/python3.8/site-packages/matplotlib/cbook/__init__.py:1376: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n  X = np.atleast_1d(X.T if isinstance(X, np.ndarray) else np.asarray(X))\n"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 600x600 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 600x600 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 600x600 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 600x600 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "execution_count": 6,
      "metadata": {}
    },
    {
      "cell_type": "markdown",
      "source": [
        "From the box plots, it looks like species 0 and 2 (Adelie and Chinstrap) have similar data profiles for culmen depth, flipper length, and body mass, but Chinstraps tend to have longer culmens. Species 1 (Gentoo) tends to have fairly clearly differentiated features from the others; which should help us train a good classification model.\n",
        "\n",
        "### Prepare the data\n",
        "\n",
        "Just as for binary classification, before training the model, we need to separate the features and label, and then split the data into subsets for training and validation. We'll also apply a *stratification* technique when splitting the data to maintain the proportion of each label value in the training and validation datasets."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "# Separate features and labels\n",
        "penguins_X, penguins_y = penguins[penguin_features].values, penguins[penguin_label].values\n",
        "\n",
        "# Split data 70%-30% into training set and test set\n",
        "x_penguin_train, x_penguin_test, y_penguin_train, y_penguin_test = train_test_split(penguins_X, penguins_y,\n",
        "                                                                                    test_size=0.30,\n",
        "                                                                                    random_state=0,\n",
        "                                                                                    stratify=penguins_y)\n",
        "\n",
        "print ('Training Set: %d, Test Set: %d \\n' % (x_penguin_train.shape[0], x_penguin_test.shape[0]))"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "Training Set: 239, Test Set: 103 \n\n"
        }
      ],
      "execution_count": 7,
      "metadata": {
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Train and evaluate a multiclass classifier\n",
        "\n",
        "Now that we have a set of training features and corresponding training labels, we can fit a multiclass classification algorithm to the data to create a model. Most scikit-learn classification algorithms inherently support multiclass classification. We'll try a logistic regression algorithm."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.linear_model import LogisticRegression\n",
        "\n",
        "# Set regularization rate\n",
        "reg = 0.1\n",
        "\n",
        "# train a logistic regression model on the training set\n",
        "multi_model = LogisticRegression(C=1/reg, solver='lbfgs', multi_class='auto', max_iter=10000).fit(x_penguin_train, y_penguin_train)\n",
        "print (multi_model)"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "LogisticRegression(C=10.0, class_weight=None, dual=False, fit_intercept=True,\n                   intercept_scaling=1, l1_ratio=None, max_iter=10000,\n                   multi_class='auto', n_jobs=None, penalty='l2',\n                   random_state=None, solver='lbfgs', tol=0.0001, verbose=0,\n                   warm_start=False)\n"
        }
      ],
      "execution_count": 8,
      "metadata": {
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now we can use the trained model to predict the labels for the test features, and compare the predicted labels to the actual labels:"
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "penguin_predictions = multi_model.predict(x_penguin_test)\n",
        "print('Predicted labels: ', penguin_predictions[:15])\n",
        "print('Actual labels   : ' ,y_penguin_test[:15])"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "Predicted labels:  [0 1 0 2 2 1 1 1 0 2 2 1 2 1 2]\nActual labels   :  [0 1 2 2 2 1 1 1 0 2 2 1 2 1 2]\n"
        }
      ],
      "execution_count": 9,
      "metadata": {
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Let's look at a classification report."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn. metrics import classification_report\n",
        "\n",
        "print(classification_report(y_penguin_test, penguin_predictions))"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "              precision    recall  f1-score   support\n\n           0       0.96      0.98      0.97        45\n           1       1.00      1.00      1.00        37\n           2       0.95      0.90      0.93        21\n\n    accuracy                           0.97       103\n   macro avg       0.97      0.96      0.96       103\nweighted avg       0.97      0.97      0.97       103\n\n"
        }
      ],
      "execution_count": 10,
      "metadata": {
        "scrolled": true,
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "As with binary classification, the report includes *precision* and *recall* metrics for each class. However, while with binary classification we could focus on the scores for the *positive* class; in this case, there are multiple classes so we need to look at an overall metric (either the macro or weighted average) to get a sense of how well the model performs across all three classes.\n",
        "\n",
        "You can get the overall metrics separately from the report using the scikit-learn metrics score classes, but with multiclass results you must specify which average metric you want to use for precision and recall."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import accuracy_score, precision_score, recall_score\n",
        "\n",
        "print(\"Overall Accuracy:\",accuracy_score(y_penguin_test, penguin_predictions))\n",
        "print(\"Overall Precision:\",precision_score(y_penguin_test, penguin_predictions, average='macro'))\n",
        "print(\"Overall Recall:\",recall_score(y_penguin_test, penguin_predictions, average='macro'))"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "Overall Accuracy: 0.970873786407767\nOverall Precision: 0.9688405797101449\nOverall Recall: 0.9608465608465608\n"
        }
      ],
      "execution_count": 11,
      "metadata": {
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now let's look at the confusion matrix for our model:"
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import confusion_matrix\n",
        "\n",
        "# Print the confusion matrix\n",
        "mcm = confusion_matrix(y_penguin_test, penguin_predictions)\n",
        "print(mcm)"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "[[44  0  1]\n [ 0 37  0]\n [ 2  0 19]]\n"
        }
      ],
      "execution_count": 12,
      "metadata": {
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "The confusion matrix shows the intersection of predicted and actual label values for each class - in simple terms, the diagonal intersections from top-left to bottom-right indicate the number of correct predictions.\n",
        "\n",
        "When dealing with multiple classes, it's generally more intuitive to visualize this as a heat map, like this:"
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "\n",
        "plt.imshow(mcm, interpolation=\"nearest\", cmap=plt.cm.Blues)\n",
        "plt.colorbar()\n",
        "tick_marks = np.arange(len(penguin_classes))\n",
        "plt.xticks(tick_marks, penguin_classes, rotation=45)\n",
        "plt.yticks(tick_marks, penguin_classes)\n",
        "plt.xlabel(\"Predicted Species\")\n",
        "plt.ylabel(\"Actual Species\")\n",
        "plt.show()"
      ],
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 640x480 with 2 Axes>",
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "execution_count": 13,
      "metadata": {}
    },
    {
      "cell_type": "markdown",
      "source": [
        "The darker squares in the confusion matrix plot indicate high numbers of cases, and you can hopefully see a diagonal line of darker squares indicating cases where the predicted and actual label are the same.\n",
        "\n",
        "In the case of a multiclass classification model, a single ROC curve showing true positive rate vs false positive rate is not possible. However, you can use the rates for each class in a One vs Rest (OVR) comparison to create a ROC chart for each class."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import roc_curve\n",
        "from sklearn.metrics import roc_auc_score\n",
        "\n",
        "# Get class probability scores\n",
        "penguin_prob = multi_model.predict_proba(x_penguin_test)\n",
        "\n",
        "# Get ROC metrics for each class\n",
        "fpr = {}\n",
        "tpr = {}\n",
        "thresh ={}\n",
        "for i in range(len(penguin_classes)):    \n",
        "    fpr[i], tpr[i], thresh[i] = roc_curve(y_penguin_test, penguin_prob[:,i], pos_label=i)\n",
        "    \n",
        "# Plot the ROC chart\n",
        "plt.plot(fpr[0], tpr[0], linestyle='--',color='orange', label=penguin_classes[0] + ' vs Rest')\n",
        "plt.plot(fpr[1], tpr[1], linestyle='--',color='green', label=penguin_classes[1] + ' vs Rest')\n",
        "plt.plot(fpr[2], tpr[2], linestyle='--',color='blue', label=penguin_classes[2] + ' vs Rest')\n",
        "plt.title('Multiclass ROC curve')\n",
        "plt.xlabel('False Positive Rate')\n",
        "plt.ylabel('True Positive rate')\n",
        "plt.legend(loc='best')\n",
        "plt.show()"
      ],
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 640x480 with 1 Axes>",
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "execution_count": 14,
      "metadata": {
        "scrolled": true
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "To quantify the ROC performance, you can calculate an aggregate area under the curve score that is averaged across all of the OVR curves."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "auc = roc_auc_score(y_penguin_test,penguin_prob, multi_class='ovr')\n",
        "print('Average AUC:', auc)"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "Average AUC: 0.9981999902100828\n"
        }
      ],
      "execution_count": 15,
      "metadata": {}
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Preprocess data in a pipeline\n",
        "\n",
        "Again, just like with binary classification, you can use a pipeline to apply preprocessing steps to the data before fitting it to an algorithm to train a model. Let's see if we can improve the penguin predictor by scaling the numeric features in a transformation steps before training. We'll also try a different algorithm (a support vector machine), just to show that we can!"
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "from sklearn.compose import ColumnTransformer\n",
        "from sklearn.pipeline import Pipeline\n",
        "from sklearn.svm import SVC\n",
        "\n",
        "# Define preprocessing for numeric columns (scale them)\n",
        "feature_columns = [0,1,2,3]\n",
        "feature_transformer = Pipeline(steps=[\n",
        "    ('scaler', StandardScaler())\n",
        "    ])\n",
        "\n",
        "# Create preprocessing steps\n",
        "preprocessor = ColumnTransformer(\n",
        "    transformers=[\n",
        "        ('preprocess', feature_transformer, feature_columns)])\n",
        "\n",
        "# Create training pipeline\n",
        "pipeline = Pipeline(steps=[('preprocessor', preprocessor),\n",
        "                           ('regressor', SVC(probability=True))])\n",
        "\n",
        "\n",
        "# fit the pipeline to train a linear regression model on the training set\n",
        "multi_model = pipeline.fit(x_penguin_train, y_penguin_train)\n",
        "print (multi_model)"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "Pipeline(memory=None,\n         steps=[('preprocessor',\n                 ColumnTransformer(n_jobs=None, remainder='drop',\n                                   sparse_threshold=0.3,\n                                   transformer_weights=None,\n                                   transformers=[('preprocess',\n                                                  Pipeline(memory=None,\n                                                           steps=[('scaler',\n                                                                   StandardScaler(copy=True,\n                                                                                  with_mean=True,\n                                                                                  with_std=True))],\n                                                           verbose=False),\n                                                  [0, 1, 2, 3])],\n                                   verbose=False)),\n                ('regressor',\n                 SVC(C=1.0, break_ties=False, cache_size=200, class_weight=None,\n                     coef0=0.0, decision_function_shape='ovr', degree=3,\n                     gamma='scale', kernel='rbf', max_iter=-1, probability=True,\n                     random_state=None, shrinking=True, tol=0.001,\n                     verbose=False))],\n         verbose=False)\n"
        }
      ],
      "execution_count": 16,
      "metadata": {
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now we can evaluate the new model."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "# Get predictions from test data\n",
        "penguin_predictions = multi_model.predict(x_penguin_test)\n",
        "penguin_prob = multi_model.predict_proba(x_penguin_test)\n",
        "\n",
        "# Overall metrics\n",
        "print(\"Overall Accuracy:\",accuracy_score(y_penguin_test, penguin_predictions))\n",
        "print(\"Overall Precision:\",precision_score(y_penguin_test, penguin_predictions, average='macro'))\n",
        "print(\"Overall Recall:\",recall_score(y_penguin_test, penguin_predictions, average='macro'))\n",
        "print('Average AUC:', roc_auc_score(y_penguin_test,penguin_prob, multi_class='ovr'))\n",
        "\n",
        "# Confusion matrix\n",
        "plt.imshow(mcm, interpolation=\"nearest\", cmap=plt.cm.Blues)\n",
        "plt.colorbar()\n",
        "tick_marks = np.arange(len(penguin_classes))\n",
        "plt.xticks(tick_marks, penguin_classes, rotation=45)\n",
        "plt.yticks(tick_marks, penguin_classes)\n",
        "plt.xlabel(\"Predicted Species\")\n",
        "plt.ylabel(\"Actual Species\")\n",
        "plt.show()"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "Overall Accuracy: 0.9805825242718447\nOverall Precision: 0.9767195767195768\nOverall Recall: 0.9767195767195768\nAverage AUC: 0.9990361381446328\n"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": "<Figure size 640x480 with 2 Axes>",
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "execution_count": 17,
      "metadata": {
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Use the model with new data observations\n",
        "\n",
        "Now let's save our trained model so we can use it again later."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "import joblib\n",
        "\n",
        "# Save the model as a pickle file\n",
        "filename = './penguin_model.pkl'\n",
        "joblib.dump(multi_model, filename)"
      ],
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 18,
          "data": {
            "text/plain": "['./penguin_model.pkl']"
          },
          "metadata": {}
        }
      ],
      "execution_count": 18,
      "metadata": {}
    },
    {
      "cell_type": "markdown",
      "source": [
        "OK, so now we have a trained model. Let's use it to predict the class of a new penguin observation:"
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "# Load the model from the file\n",
        "multi_model = joblib.load(filename)\n",
        "\n",
        "# The model accepts an array of feature arrays (so you can predict the classes of multiple penguin observations in a single call)\n",
        "# We'll create an array with a single array of features, representing one penguin\n",
        "x_new = np.array([[50.4,15.3,224,5550]])\n",
        "print ('New sample: {}'.format(x_new[0]))\n",
        "\n",
        "# The model returns an array of predictions - one for each set of features submitted\n",
        "# In our case, we only submitted one penguin, so our prediction is the first one in the resulting array.\n",
        "penguin_pred = multi_model.predict(x_new)[0]\n",
        "print('Predicted class is', penguin_classes[penguin_pred])"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "New sample: [  50.4   15.3  224.  5550. ]\nPredicted class is Gentoo\n"
        }
      ],
      "execution_count": 19,
      "metadata": {
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "You can also submit a batch of penguin observations to the model, and get back a prediction for each one."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "# This time our input is an array of two feature arrays\n",
        "x_new = np.array([[49.5,18.4,195, 3600],\n",
        "         [38.2,20.1,190,3900]])\n",
        "print ('New samples:\\n{}'.format(x_new))\n",
        "\n",
        "# Call the web service, passing the input data\n",
        "predictions = multi_model.predict(x_new)\n",
        "\n",
        "# Get the predicted classes.\n",
        "for prediction in predictions:\n",
        "    print(prediction, '(' + penguin_classes[prediction] +')')"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": "New samples:\n[[  49.5   18.4  195.  3600. ]\n [  38.2   20.1  190.  3900. ]]\n2 (Chinstrap)\n0 (Adelie)\n"
        }
      ],
      "execution_count": 20,
      "metadata": {
        "tags": []
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Summary\n",
        "\n",
        "Classification is one of the most common forms of machine learning, and by following the basic principles we've discussed in this notebook you should be able to train and evaluate classification models with scikit-learn. It's worth spending some time investigating classification algorithms in more depth, and a good starting point is the [Scikit-Learn documentation](https://scikit-learn.org/stable/user_guide.html)."
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [],
      "outputs": [],
      "execution_count": null,
      "metadata": {}
    }
  ],
  "metadata": {
    "kernelspec": {
      "name": "conda-env-azureml_py38-py",
      "language": "python",
      "display_name": "azureml_py38"
    },
    "language_info": {
      "name": "python",
      "version": "3.8.5",
      "mimetype": "text/x-python",
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "pygments_lexer": "ipython3",
      "nbconvert_exporter": "python",
      "file_extension": ".py"
    },
    "kernel_info": {
      "name": "conda-env-azureml_py38-py"
    },
    "nteract": {
      "version": "nteract-front-end@1.0.0"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 2
}