B_modelFakeReal.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# **Modeling - Real vs Fake News**\n",
        "\n",
        "<small>Written by: Ali Tobah - tobah@umich.edu</small>\n",
        "\n",
        "**A note on images:**\n",
        "\n",
        "Charts generated by Altair don't show up in GitHub. So PNG versions of some of the charts in this notebook are also shown embedded in Markdown cells following the code that generates them. In the **Appendix**, there is a link to an explanation of why this happens, and another on the workaround used here. Charts generated by Matplotlib show up normally."
      ],
      "metadata": {
        "id": "YgbkCIXuMfk_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Uncomment to install as needed\n",
        "#!pip install empath\n",
        "#!pip install scattertext\n",
        "\n",
        "# Recommended by Scattertext authors,\n",
        "# \"in order to take full advantage of Scattertext\"\n",
        "# but might not need to install; might be\n",
        "# installed automatically.\n",
        "#!pip install jieba\n",
        "#!pip install spacy\n",
        "#!pip install astropy\n",
        "#!pip install flashtext\n",
        "#!pip install gensim\n",
        "#!pip install umap-learn"
      ],
      "metadata": {
        "id": "JmY3u6c11jH7"
      },
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import altair as alt\n",
        "import matplotlib.pyplot as plt\n",
        "from scipy import sparse\n",
        "from sklearn import preprocessing\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.feature_extraction.text import TfidfVectorizer\n",
        "from collections import Counter\n",
        "from sklearn.dummy import DummyClassifier\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.svm import LinearSVC\n",
        "from sklearn.metrics import f1_score, confusion_matrix, ConfusionMatrixDisplay\n",
        "from empath import Empath\n",
        "lexicon = Empath()\n",
        "import scattertext as st\n",
        "import warnings\n",
        "warnings.filterwarnings('ignore')"
      ],
      "metadata": {
        "id": "ycWkglbQm_ZU"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Constants\n",
        "RED = '#d62728'    # From Vega color schemes category10\n",
        "GREEN = '#4daf4a'    # From Vega color schemes set1\n",
        "RANDOM_STATE = 0\n",
        "dirPath = 'data/processed'"
      ],
      "metadata": {
        "id": "HinnK4HWkRH2"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **Function definitions**"
      ],
      "metadata": {
        "id": "CH0CREWGB51s"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def splitStratify(originalDF, stratify=False):\n",
        "    '''\n",
        "    Splits the data while stratifying on domain if applicable\n",
        "\n",
        "    Called by: modelData()\n",
        "    Calls: None\n",
        "\n",
        "    Args:\n",
        "      - Original data as a dataframe\n",
        "      - Boolean indicating whether to stratify by domain\n",
        "    Returns:\n",
        "      - Training df\n",
        "      - Testing df\n",
        "      - Chart of training data by label and domain\n",
        "      \n",
        "    Notes:\n",
        "      - Prints training and test dataframe shapes\n",
        "      - Prints fake (1) / real (0) label counts in the training data\n",
        "  '''\n",
        "\n",
        "    dataDF = originalDF.copy()\n",
        "\n",
        "    # Split with stratification on Domain if applicable\n",
        "    if stratify == False:\n",
        "        trainDF, testDF = train_test_split(dataDF, test_size=0.2, random_state=RANDOM_STATE)\n",
        "    else:\n",
        "        trainDF, testDF = train_test_split(dataDF, test_size=0.2,\n",
        "                                      random_state=RANDOM_STATE, stratify=dataDF[['Domain']])\n",
        "\n",
        "    # Check the sizes\n",
        "    print(\"Training and test dataframe shapes:\", trainDF.shape, testDF.shape)\n",
        "\n",
        "    # Check balance of labels in the training data\n",
        "    print('Fake (1) / real (0) label counts in the training data:', Counter(trainDF['intLabel']))\n",
        "\n",
        "    print('\\n')\n",
        "\n",
        "    # This is admittedly a kluge, but still better than the\n",
        "    # alternative: Convert labels back to text in the training\n",
        "    # df in a new column, just for the chart, then drop it.\n",
        "    trainDF['Labels'] = trainDF['intLabel'].map({1: 'fake', 0: 'real'})\n",
        "\n",
        "    domainColor = ['fake', 'real']\n",
        "    rangeColor = [RED, GREEN]\n",
        "\n",
        "    # Do a visual check\n",
        "    trainChartDF = trainDF.groupby(['Domain', 'Labels']).count().rename(columns={'Text': 'Number of Articles'}).reset_index()\n",
        "    trainChart = alt.Chart(trainChartDF).mark_bar().encode(\n",
        "        x = alt.X('Number of Articles:Q'),\n",
        "        y = alt.Y('Labels:N', title=None, axis=alt.Axis(labels=False, tickSize=0)),\n",
        "        color = alt.Color('Labels:N', scale=alt.Scale(domain=domainColor, range=rangeColor)),\n",
        "        row = alt.Row('Domain:N')\n",
        "        ).properties(\n",
        "            title = 'Training data by domain',\n",
        "            height=25,\n",
        "            width=300\n",
        "        ).configure_axis(\n",
        "            labelFontSize=10,\n",
        "            titleFontSize=12\n",
        "        ).configure_title(\n",
        "            fontSize=15\n",
        "        ).configure_legend(\n",
        "            titleFontSize=12\n",
        "        )\n",
        "    \n",
        "    trainDF = trainDF.drop(['Labels'], axis=1)\n",
        "\n",
        "    return trainDF, testDF, trainChart"
      ],
      "metadata": {
        "id": "U-DgKjiUyL_g"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def plot_confusion_matrix(y_test=None, y_pred=None, labels=None):\n",
        "    \"\"\"\n",
        "    Plots the confusion matrix for the given model results (in line plot)\n",
        "\n",
        "    Called by: getF1Scores()\n",
        "    Calls: None\n",
        "\n",
        "    Args:\n",
        "        y_test (list or series): Real labels for the test dataset\n",
        "        y_pred (list or series): Predicted labels for the test dataset\n",
        "        labels (list): Classification label names\n",
        "    Returns:\n",
        "        Confusion matrix plot\n",
        "    \n",
        "    Notes: Slightly revised from CGregg's code\n",
        "    \"\"\"\n",
        "    conf_matrix = confusion_matrix(y_test, y_pred, labels=labels, normalize='all')\n",
        "    cm_disp = ConfusionMatrixDisplay(confusion_matrix=conf_matrix,\n",
        "                                    display_labels=['real', 'fake']\n",
        "                                    )\n",
        "\n",
        "    confPlot = cm_disp.plot(cmap='Purples')\n",
        "\n",
        "    return confPlot"
      ],
      "metadata": {
        "id": "v2oVGqN93xJO"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Prepare a text vectorizer and train\n",
        "# logistic regression and linearSVC models\n",
        "def getF1Scores(trainDFVar, testDFVar, stopWords=None, min_df=1, ngrams=1):\n",
        "    '''\n",
        "    Prepares the vectorizer, trains the models and returns F1 scores and accuracy\n",
        "\n",
        "    Called by: modelData()\n",
        "    Calls: plot_confusion_matrix()\n",
        "\n",
        "    Args:\n",
        "      - Training df\n",
        "      - Test df\n",
        "      - StopWords\n",
        "      - Minimum document frequency for TfidfVectorizer\n",
        "      - Maximum ngrams for TfidfVectorizer\n",
        "    Returns:\n",
        "      - Text vectorizer\n",
        "      - Xtrain array after fit and transform with vectorizer, addition of extra features\n",
        "      - Trained Logistic Regression model\n",
        "      - Trained LinearSVC model\n",
        "      - The confusion matrix for the Logistic Regression model\n",
        "      \n",
        "    Notes:\n",
        "      - Prints F1 scores and accuracy for a dummy (random) classifier, a Logistic Regression model and a LinearSVC model\n",
        "      - Mostly default parameters for vectorizer and models\n",
        "    '''\n",
        "\n",
        "    # Vectorize the training text\n",
        "    textVec = TfidfVectorizer(stop_words=stopWords, min_df=min_df, ngram_range=(1,ngrams))\n",
        "    Xtrain = textVec.fit_transform(trainDFVar.Text)\n",
        "    \n",
        "    # If there are features other than the vectorized text, scale\n",
        "    # them and convert to sparse array (same type as Xtrain),\n",
        "    # then merge with Xtrain\n",
        "    tempDF = trainDFVar.drop(['Text', 'Domain', 'intLabel'], axis=1)\n",
        "    if tempDF.shape[1] != 0:\n",
        "        tempArray = sparse.csr_matrix(tempDF.values)\n",
        "        tempScaler = preprocessing.MaxAbsScaler().fit(tempArray)\n",
        "        tempArray = tempScaler.transform(tempArray)\n",
        "    \n",
        "        # Merge the two sparse arrays\n",
        "        Xtrain = sparse.hstack([Xtrain, tempArray])\n",
        "\n",
        "\n",
        "\n",
        "    # Repeat for the test data\n",
        "    Xtest = textVec.transform(testDFVar.Text)\n",
        "\n",
        "    tempDF = testDFVar.drop(['Text', 'Domain', 'intLabel'], axis=1)\n",
        "    if tempDF.shape[1] != 0:\n",
        "        tempArray = sparse.csr_matrix(tempDF.values)\n",
        "        tempArray = tempScaler.transform(tempArray)\n",
        "\n",
        "        Xtest = sparse.hstack([Xtest, tempArray])\n",
        "\n",
        "\n",
        "    # Get the labels\n",
        "    ytrain = list(trainDFVar.intLabel)\n",
        "    ytest = list(testDFVar.intLabel)\n",
        "    \n",
        "    # Train a dummy classifier for comparison\n",
        "    clfRandom = DummyClassifier(strategy=\"stratified\", random_state=RANDOM_STATE)\n",
        "    clfRandom.fit(Xtrain, ytrain)\n",
        "\n",
        "    # Train a logistic regression classifier\n",
        "    # and an SVC classifier\n",
        "\n",
        "    clfLR = LogisticRegression(random_state=RANDOM_STATE)\n",
        "    clfLR.fit(Xtrain, ytrain)\n",
        "\n",
        "    clfSVC = LinearSVC(loss='hinge', C=0.4, random_state=RANDOM_STATE)\n",
        "    clfSVC.fit(Xtrain, ytrain)\n",
        "\n",
        "    # Get predictions and F1 scores\n",
        "\n",
        "    predRandom = clfRandom.predict(Xtest)\n",
        "    predLR = clfLR.predict(Xtest)\n",
        "    predSVC = clfSVC.predict(Xtest)\n",
        "\n",
        "    f1Rand = f1_score(ytest, predRandom)\n",
        "    accRand = clfRandom.score(Xtest, ytest)\n",
        "    print('Random f1 score:', f1Rand, '...', 'Accuracy:', accRand)\n",
        "    f1LR = f1_score(ytest, predLR)\n",
        "    accLR = clfLR.score(Xtest, ytest)\n",
        "    print('Logistic regression f1 score:', f1LR, '...', 'Accuracy:', accLR)\n",
        "    f1SVC = f1_score(ytest, predSVC)\n",
        "    accSVC = clfSVC.score(Xtest, ytest)\n",
        "    print('SVC f1 score:', f1SVC, '...', 'Accuracy:', accSVC)\n",
        "\n",
        "    print('\\n')\n",
        "\n",
        "    # Create the confusion matrix for the Logistic Regression model\n",
        "    confMtxPlot = plot_confusion_matrix(ytest, predLR, clfLR.classes_)\n",
        "\n",
        "    # Return the vectorizer, the models for use\n",
        "    # in the explanation and the confusion matrix\n",
        "    return textVec, Xtrain, clfLR, clfSVC, confMtxPlot\n"
      ],
      "metadata": {
        "id": "bzpoytTpI24I"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Explain the predictions\n",
        "def explainPred(featList, trainArray, clfLR, clfSVC):\n",
        "    '''\n",
        "    Plots highest coefficients\n",
        "\n",
        "    Called by: modelData()\n",
        "    Calls: None\n",
        "    \n",
        "    Args:\n",
        "       - List of feature names\n",
        "       - Transformed training data array\n",
        "       - Trained Logistic Regression model\n",
        "       - Trained LinearSVC model\n",
        "    Returns:\n",
        "      - Horizontally concatenated chart showing the weights for each model for the 20 most important features\n",
        "    '''\n",
        "\n",
        "    domainColor = ['fake', 'real']\n",
        "    rangeColor = [RED, GREEN]\n",
        "\n",
        "    # Standardized list of Logistic Regression model coefficients\n",
        "    stdCoefListLR = list((np.std(trainArray.toarray(), 0) * clfLR.coef_)[0])\n",
        "\n",
        "    # Create dataframe with feature names and get top 20\n",
        "    LRfeatCoefDF = pd.DataFrame({'Features': featList, 'LRWeights': stdCoefListLR})\n",
        "    LRfeatCoefDF = LRfeatCoefDF.sort_values(by='LRWeights', key=abs, ascending=False)\n",
        "    LRfeatCoefDF20 = LRfeatCoefDF.head(20)\n",
        "    # Add a label column to get the legend\n",
        "    LRfeatCoefDF20['Labels'] = LRfeatCoefDF20['LRWeights'].apply(lambda x: 'fake' if x > 0 else 'real')\n",
        "\n",
        "    LRChart = alt.Chart(LRfeatCoefDF20).mark_bar().encode(\n",
        "        x = alt.X(\"LRWeights:Q\"),\n",
        "        y = alt.Y(\"Features:N\", sort=None, axis=alt.Axis(grid=True)),\n",
        "        color = alt.Color('Labels', scale=alt.Scale(domain=domainColor, range=rangeColor))\n",
        "    ).properties(width=250, height=300)\n",
        "\n",
        "    # Standardized list of LinearSVC model coefficients\n",
        "    stdCoefListSVC = list((np.std(trainArray.toarray(), 0) * clfSVC.coef_)[0])\n",
        "\n",
        "    # Create dataframe with feature names and get top 20\n",
        "    SVCfeatCoefDF = pd.DataFrame({'Features': featList, 'SVCWeights': stdCoefListSVC})\n",
        "    SVCfeatCoefDF = SVCfeatCoefDF.sort_values(by='SVCWeights', key=abs, ascending=False)\n",
        "    SVCfeatCoefDF20 = SVCfeatCoefDF.head(20)\n",
        "\n",
        "    # Did not add a labels column here, and coded\n",
        "    # the color using a condition instead. Just a\n",
        "    # different way of doing it.\n",
        "    SVCChart = alt.Chart(SVCfeatCoefDF20).mark_bar().encode(\n",
        "        x = alt.X(\"SVCWeights:Q\"),\n",
        "        y = alt.Y(\"Features:N\", sort=None, axis=alt.Axis(grid=True)),\n",
        "        color = alt.condition(\n",
        "            alt.datum.SVCWeights > 0,\n",
        "            alt.value(RED),  # The positive (fake) color\n",
        "            alt.value(GREEN)  # The negative (real) color\n",
        "        )\n",
        "    ).properties(\n",
        "        width=250, height=300\n",
        "    )\n",
        "\n",
        "    finalChart = (LRChart | SVCChart).properties(\n",
        "            title = 'Feature Weights',\n",
        "        ).configure_axis(\n",
        "            labelFontSize=10,\n",
        "            titleFontSize=12\n",
        "        ).configure_title(\n",
        "            fontSize=15\n",
        "        )\n",
        "        \n",
        "    return finalChart"
      ],
      "metadata": {
        "id": "frmsLt47BHD6"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def modelData(dataDF, stratify=False):\n",
        "    '''\n",
        "    Splits the data, runs the models and plots the results (i.e. run the functions above in sequence)\n",
        "\n",
        "    Called by: Main code\n",
        "    Calls:\n",
        "      - splitStratify()\n",
        "      - getF1Scores()\n",
        "      - explainPred()\n",
        "      \n",
        "    Args:\n",
        "      - Original data as a dataframe\n",
        "      - Boolean indicating whether to stratify by domain\n",
        "    Returns:\n",
        "      - Training data chart from splitStratify()\n",
        "      - Feature weights chart from explainPred()\n",
        "      - Confusion matrix for the Logistic Regression model\n",
        "    '''\n",
        "  \n",
        "    # Split the data with stratification if applicable\n",
        "    trainDF, testDF, trainChart = splitStratify(articlesDF, stratify=stratify)\n",
        "\n",
        "    # Get prepare a text vectorizer, train two models\n",
        "    # and get F1 scores\n",
        "    textVectorizer, trainArray, clfLogReg, clfSVC, confMtxPlot = getF1Scores(trainDF, testDF,\n",
        "                                                            stopWords='english', min_df=75, ngrams=2)\n",
        "    \n",
        "    # Get the text features\n",
        "    featNamesList = list((textVectorizer.get_feature_names_out()))\n",
        "    # Include additional features from the articles df if any\n",
        "    addnlFeats = list(articlesDF.columns)[3:]   # The first 3 columns are the text, domains and labels\n",
        "    featNamesList.extend(addnlFeats)\n",
        "\n",
        "    # Explain and plot the predictions\n",
        "    predChart = explainPred(featNamesList, trainArray, clfLogReg, clfSVC)\n",
        "\n",
        "    return trainChart, predChart, confMtxPlot"
      ],
      "metadata": {
        "id": "43xxYRailuZm"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **Load the data**"
      ],
      "metadata": {
        "id": "wOXNKmTm0Zn3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "inFile = dirPath + '/allnewsdataFakeReal.csv'\n",
        "articlesDF = pd.read_csv(inFile, index_col=0)\n",
        "\n",
        "# Add integer labels and drop the text labels column\n",
        "# Note that the authors of the dataset use 'legit' to\n",
        "# imply real news. We use 'real' from this point on.\n",
        "articlesDF['intLabel'] = articlesDF['Label'].map({'fake': 1, 'legit': 0})\n",
        "articlesDF = articlesDF.drop(['Label'], axis=1)\n",
        "\n",
        "# Uncomment to use only celebrity data.\n",
        "# Can limit to all the other data by using '!='\n",
        "#articlesDF = articlesDF[articlesDF['Domain'] == 'celebrity']\n",
        "\n",
        "\n",
        "print(articlesDF.shape)\n",
        "articlesDF.head()"
      ],
      "metadata": {
        "id": "Bsy5Rhvzzp8C",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 225
        },
        "outputId": "e998ed26-dc5e-4a99-a7c3-66251bb92e50"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(980, 3)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                Text Domain  intLabel\n",
              "0  Supreme Court Forces Schools Into Socialism\\n\\...    edu         1\n",
              "1  Brexit fist lost\\nThe oil-rich Gulf country Qa...    biz         1\n",
              "2  Leftist universities seek to put women in posi...    edu         1\n",
              "3  New intelligence suggests ISIS rigging laptop ...   tech         1\n",
              "4  Microsoft Aims to spread liberalism on the suc...    edu         1"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-9886c0f4-fd33-45fe-bf27-54279145f9c4\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>Text</th>\n",
              "      <th>Domain</th>\n",
              "      <th>intLabel</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Supreme Court Forces Schools Into Socialism\\n\\...</td>\n",
              "      <td>edu</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Brexit fist lost\\nThe oil-rich Gulf country Qa...</td>\n",
              "      <td>biz</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Leftist universities seek to put women in posi...</td>\n",
              "      <td>edu</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>New intelligence suggests ISIS rigging laptop ...</td>\n",
              "      <td>tech</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Microsoft Aims to spread liberalism on the suc...</td>\n",
              "      <td>edu</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-9886c0f4-fd33-45fe-bf27-54279145f9c4')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-9886c0f4-fd33-45fe-bf27-54279145f9c4 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-9886c0f4-fd33-45fe-bf27-54279145f9c4');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **Modeling**"
      ],
      "metadata": {
        "id": "wvjBFsOjl-5J"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "####Features: Baseline. Text only, no additional features."
      ],
      "metadata": {
        "id": "V23xuIgCDUho"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# NOTE: The two Altair charts 'trainChart' and 'predChart' will\n",
        "# not show on GitHub as explained in the intro and the appendix.\n",
        "# They are embedded in following Markdown cell as a workaround.\n",
        "# The third chart is generated by Matplotlib and isn't affected.\n",
        "\n",
        "trainChart, predChart, confMtxPlot = modelData(articlesDF, stratify=False)\n",
        "# Display the training data split\n",
        "#trainChart.display()\n",
        "\n",
        "# Display highest weighted coefficients\n",
        "#predChart.display()\n",
        "\n",
        "# Show the confusion matrix\n",
        "confMtxPlot;"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 444
        },
        "id": "BShSGJWInpTO",
        "outputId": "396f329b-92f5-4c08-ba29-02412ed0b2af"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training and test dataframe shapes: (784, 3) (196, 3)\n",
            "Fake (1) / real (0) label counts in the training data: Counter({1: 405, 0: 379})\n",
            "\n",
            "\n",
            "Random f1 score: 0.5340909090909091 ... Accuracy: 0.5816326530612245\n",
            "Logistic regression f1 score: 0.6774193548387096 ... Accuracy: 0.6938775510204082\n",
            "SVC f1 score: 0.6703296703296703 ... Accuracy: 0.6938775510204082\n",
            "\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUoAAAEGCAYAAAADs9wSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAeR0lEQVR4nO3deZTU1Z338fenG5CIohBQw6KAoohoEAG3ExdEgkvEPGrUaE4cdYhb8InjGDPjaEJOEpckTkwwigkTdxJjokRB4ihGjajgLlsEZOeJbIosKt39ff6oX2N1011VLV2/7io+L08dfvu9Rcun72+7VxGBmZk1rqKlK2Bm1to5KM3M8nBQmpnl4aA0M8vDQWlmlkeblq5Ac2irnaM9u7d0NawJehzYpaWrYE00d85bqyOi6/Yco7P2iy1sKmjbDaycGhEjt6e85lIWQdme3TmMi1u6GtYEt9wzuqWrYE00dEjPxdt7jC1sKvjf6t/4Yav5bVoWQWlmpUNSYRu2oke8HZRmlhoBqiwwKKuKWpUmcVCaWXoEhTYoWxMHpZmlqwST0kFpZqkqwZx0UJpZmoQqSi8pHZRmlh5Rkk1KB6WZpUZAhVuUZmZ5lF5OOijNLEXC1yjNzPIpwUuUDkozS1kJJqWD0szSI1FR6CuMrYiD0szSVYItSnfca2apqX2MspBP3mNJIyXNkzRf0rUNrL9E0luSXpf0vKT+yfJekjYny1+XdEe+styiNLNUFdzNWu5jVALjgBOBZcAMSZMiYnbWZg9ExB3J9qcBPwdqOwJeEBEDCy3PLUozS5cK/OQ2FJgfEQsj4hNgIjAqe4OIWJ8124Ht6OHSLUozS0/TnqPsImlm1vz4iBifTHcHlmatWwYcvk1x0uXAVUA7YFjWqt6SXgPWA9dFxHO5KuKgNLNUNSEoV0fE4O0pKyLGAeMkfR24DvgmsBLYOyLWSDoMeETSQfVaoHX41NvMUiSkwj55LAd6Zs33SJY1ZiJwOkBEfBwRa5LpV4AFwP65CnNQmll6RCZ1CvnkNgPoK6m3pHbAOcCkOkVJfbNmTwHeSZZ3TW4GIakP0BdYmKswn3qbWWoyjwdt/13viKiSdAUwFagEJkTELEljgZkRMQm4QtJwYAuwjsxpN8AxwFhJW4Aa4JKIWJurPAelmaWquZ43j4jJwOR6y67Pmr6ykf0eBh5uSlkOSjNLj3sPMjPLz0FpZpZPCb7r7aA0s9SU6JA5DkozS1Fhz0i2Og5KM0tXCT697aA0s1RVVJReUjoozSw9ApVeTjoozSxlvkZpZtY43/U2M8vHb+aYmeVT4IA4rYyD0sxS5eFqzcxyKdGLlA5KM0tVCeakg9LM0iN8M8fMLL/Sy0kHpZmlSKKisvRezXFQmlmqfI3SzCyfEkxKB6WZpcdv5piZ5Vaij1E6KM0sZSWYlA5KM0uP/AqjmVke7hTDzCyvEsxJB2VLGTJiPy7/+UgqKiuYPOFVJt7yfJ31p44ezKhLh1BTHWze8Am3XvoXFs9ZxQnnHszX/u3ordv1OXhPLhl6Jwve+H9pf4UdztwXljLpZy9QUxMMHdWPYRcMrLN+4asrmfTzF1g5fy3n/egEDjmhz9Z1j//yJeY8vwSA4RcNYuCIfVOte6vhu97FIWkRMDgiVrd0XZpLRYUYc9vJXHPSvaxatp7bX/xXpj82j8VzVm3d5ukH3+Kx8TMBOPLUA7jkli/zvVPv46kH3+KpB98CoPeAPRj7x3Mckimoqa7hzzc/z+hfncJue3bgtm/+mYOO2Yc9+3Taus3ue+3C1244jr/d92adfec8v4Tlc1fznfvPoHpLNb/+1mP0O6on7Xdpl/bXaB1KsEmZ6rtEyii995eaWb+h3Vm+YC0r311H1ZZqpv3+bY76ygF1ttn04cdbp9t3aAsR2xxn2NkHM+0Pbxe9vgZLZq2iS8/d+HyPjrRpW8nAE/dl1t8W1dmmc7dd6db389uMW/3Pd9fR+9AvUNmmgnafa8sX+nZm3vSlKda+9ah9PKiQT2tS9NCS1EvSPEn3AG8D/yVphqQ3Jf0ga7tHJL0iaZak0cWuV0vq0q0jq5at3zq/avl6unTvuM12oy4dwr1zxzD6Jyfyq+9M2Wb9cWcdxNO/d1CmYf2qjey+Z4et87vt2YEPVm0saN8v9P0886Yv5ZOPqtj4/kcsmLmC9/9Z2L5lR0KVFQV9WpO0Tr37At8EOgJnAkPJ/HKZJOmYiHgWuDAi1kr6HDBD0sMRsaaxAyZhOhpgJ3Yr+hdoCY/+egaP/noGw845mPP/4xhuuvCRrev6De3OR5u3sGjWey1YQyvEAUf0YNns9/jVhY+yS6f27HPwnlSU4HW65tLaWouFSCu2F0fEi8CI5PMa8CrQj0yIAoyR9AbwItAza3mDImJ8RAyOiMFt2bl4NS+C1SvW07XHpy3Irt07snr5+ka3n/b7tznqtH51lh3/tQFMm+jWZFo6du1QpxX4wT83slvXDjn2qOuECwdx1QNnMHrcKQRBl33K85d7IVShgj6tSVpBWft/mICfRMTA5LNfRPxW0nHAcODIiPgimSBtn1LdUjd3xgq67/d59uq1O23aVnL82QN44bF5dbbpvl/nrdNHnNyX5fPXbp2XxHFnHuTrkynq2b8rq5d8wNrl66naUs3rTy6g/zH7FLRvTXUNG9//CIAV76xh5Ttr2f/wHsWsbuulzP+/hXxak7Tvek8Ffijp/ojYIKk7sAXYDVgXEZsk9QOOSLleqaqpruGXV07mpse/QUWlmPK711g8exUX3HA8815ZwfTH5nH6ZUMZNKwPVVU1bFi3mZsu/PPW/Q/50j68t2w9K99d14LfYsdS2aaC0685mrvGTKGmuoahpx3AXvt2ZuodM+lxYBcOOrYXS2e9x93XPMmm9R8z5/nF/PXOV7j6D2dRXVXD7aMnAdC+QzvOHXs8lW1a1zW4VDVTBkoaCfwCqAR+ExE31lt/CXA5UA1sAEZHxOxk3feAi5J1YyJias6yooG7qc1JUi/gsYgYkMxfCVycrN4AnA8sAx4BegHzgN2B70fEM4U8HrSrusVhWw9ppeCWGWV9v64sDR3S85WIGLw9x9izc58458s/Lmjb2yae22h5kiqBfwAnksmPGcC5tUGYbNMxItYn06cBl0XESEn9gQfJ3CvpBvwvsH9EVDdWl6K3KCNiETAga/4XZH4L1HdSI/v3KkrFzCx9EjTP9cehwPyIWJg5rCYCo4CtQVkbkokOQG2rcBQwMSI+Bt6VND853vTGCmv1D5ybWXlpwuXHLpJmZs2Pj4jxyXR3IPth1GXA4duWpcuBq4B2wLCsfV+st2/3XBVxUJpZqppwo2b19p7qR8Q4YJykrwPXkXlMsckclGaWHtFcp97LyTxGWKtHsqwxE4Fff8Z9032F0cysmV5hnAH0ldRbUjvgHGBS3XKU/Sz2KcA7yfQk4BxJO0nqTeaZ7ZdzFeYWpZmlRtAsrydGRJWkK8g8clgJTIiIWZLGAjMjYhJwhaThZB5BXEdy2p1s9wcyN36qgMtz3fEGB6WZpakZHyaPiMnA5HrLrs+avjLHvj8CflRoWQ5KM0tVKfYf5qA0s1S1ttcTC+GgNLN0OSjNzHKQT73NzHJqrrveaXNQmll6WmEXaoVwUJpZqkowJx2UZpayVtZ7eSEclGaWKp96m5nlIlClg9LMrFGZcb0dlGZmjVPrG2GxEA5KM0tX6eWkg9LM0uVTbzOzPHzqbWaWixyUZmY5+a63mVkBSjAnHZRmli4HpZlZLuXWe5CkXwLR2PqIGFOUGplZ2RJQUWY3c2amVgsz22GUYIOy8aCMiLuz5yXtHBGbil8lMytnpXjqnbdPdklHSpoNzE3mvyjp9qLXzMzKjzItykI+rUkhg1f8N/BlYA1ARLwBHFPMSplZ+VKB/7UmBd31joil9ZrL1cWpjpmVs8wD5y1di6YrJCiXSjoKCEltgSuBOcWtlpmVq1K8613IqfclwOVAd2AFMDCZNzNrslK8Rpm3RRkRq4HzUqiLmZW71piCBSjkrncfSX+RtErSe5IeldQnjcqZWfkpxRZlIafeDwB/AL4AdAMeAh4sZqXMrDzV9h5UyKc1KSQod46IeyOiKvncB7QvdsXMrDyVYosy17venZPJKZKuBSaSeff7bGByCnUzs3Kj0rzrnetmzitkgrH2W30ra10A3ytWpcysfDVXTEoaCfwCqAR+ExE31lt/FXAxUAWsAi6MiMXJumrgrWTTJRFxWq6ycr3r3fszfwMzswY0Vw/nkiqBccCJwDJghqRJETE7a7PXgMERsUnSpcDNZM6IATZHxMBCyyvozRxJA4D+ZF2bjIh7Ci3EzKxWM11/HArMj4iFmWNqIjAK2BqUETEta/sXgfM/a2F5g1LSDcBxZIJyMnAS8DzgoDSzJmtCi7KLpOzuHsdHxPhkujuwNGvdMuDwHMe6CJiSNd8+OXYVcGNEPJKrIoW0KM8Evgi8FhH/ImlP4L4C9jMzq0tqys2c1RExePuL1PnAYODYrMX7RMTy5JnwpyW9FRELGjtGIY8HbY6IGqBKUkfgPaDn9lTczHZMtZ1iNMPjQcupm0M9kmV1y5OGA/8JnBYRH9cuj4jlyZ8LgWeAQ3MVVkhQzpS0O3AXmTvhrwLTC9jPzGwbzRSUM4C+knpLagecA0yqW44OBe4kE5LvZS3vJGmnZLoLcDRZ1zYbUsi73pclk3dIegLoGBFv5v0aZmYNaI673hFRJekKYCqZx4MmRMQsSWOBmRExCbgF2AV4KCmz9jGgA4E7JdWQaSzeWO9u+TZyPXA+KNe6iHi1id/NzKzZ3rqJiMnUe/klIq7Pmh7eyH4vAAc3paxcLcqf5aojMKwpBRXTAYO68dRL32/palgTXHzCb1u6CtYSVJpj5uR64Pz4NCtiZuVPgMrsFUYzs2ZXVi1KM7NiKMGcdFCaWYpaYV+ThSikh3NJOl/S9cn83pKGFr9qZlZumvGB81QV8sD57cCRwLnJ/Idkeu0wM2uyUuzhvJBT78MjYpCk1wAiYl3yJLyZWdOUYce9tbYkfb8FgKSuQE1Ra2VmZau1tRYLUUhQ3gb8GdhD0o/I9CZ0XVFrZWZlqWyfo4yI+yW9ApxA5nueHhFzil4zMytLJdigLKjj3r2BTcBfspdFxJJiVszMylArvFFTiEJOvR/n00HG2gO9gXnAQUWsl5mVqbK8mRMRdXrZSHoVuqyRzc3MGtVcg4ulrclv5kTEq5JyjU1hZtaoEszJgq5RXpU1WwEMAlYUrUZmVr5qX80pMYW0KHfNmq4ic83y4eJUx8zKXdmdeicPmu8aEVenVB8zK3MlmJM5h4Jok4xLcXSaFTKz8iWJisrSS8pcLcqXyVyPfF3SJOAhYGPtyoj4U5HrZmZlqOxOvRPtgTVkxsipfZ4yAAelmTVZuQXlHskd77f5NCBrRVFrZWZlqwRzMmdQVpIZE7ehr+WgNLPPpNxalCsjYmxqNTGzsqdyG66WhluSZmbbQWX3rvcJqdXCzHYYZdWijIi1aVbEzHYMJZiTHq7WzNIjlWkP52ZmzcktSjOzPMrqGqWZWTE4KM3MclFpnnpXtHQFzGzHITKDixXyyXssaaSkeZLmS7q2gfVXSZot6U1JT0naJ2vdNyW9k3y+ma8sB6WZpUoq7JP7GKoExgEnAf2BcyX1r7fZa8DgiDgE+CNwc7JvZ+AG4HBgKHCDpE65ynNQmlmqmqlFORSYHxELI+ITYCIwKnuDiJgWEZuS2ReBHsn0l4EnI2JtRKwDngRG5irM1yjNLD1q0nC1XSTNzJofHxHjk+nuwNKsdcvItBAbcxEwJce+3XNVxEFpZqlp4thiqyNi8HaXKZ0PDAaO/azH8Km3maWqOa5RAsuBnlnzPZJl9crScOA/gdMi4uOm7JvNQWlmqVKB/+UxA+grqbekdsA5wKQ65UiHAneSCcn3slZNBUZI6pTcxBmRLGuUT73NLFXN8RxlMvDhFWQCrhKYEBGzJI0FZkbEJOAWMp2PP5TcHFoSEadFxFpJPyQTtgBj83UC5KA0s9SoaTdzcoqIycDkesuuz5oenmPfCcCEQstyUJpZigp7mLy1cVCaWapKMCcdlGaWLrcozczyKb2cdFCaWYrKcBRGM7NmJZrvrneaHJRmlqrSi0kHpZmlzKfeZmZ5lGBOOihbystT32HcVU9QU13DyRcO4txrvlRn/ZvPLWLcVU+w8K1/ct39Z3LsGQdtXXftKfcy+6VlDDh6b3786HlpV32HddCQ7px7+RFUVFTw3OR5TJn4Zp31x57aj+NHHUhNTfDx5i3cc+vfWbn4ffof1o0zLh5CZZsKqqtqeOjOl5n7+soW+hYtq9Dey1ubogWlpDHApcCrEbHNv2ZJF5DpffiKYtWhtaquruG2MZO5eco36NqjI5cdcRdHnnoAvfrvsXWbPXruxjW/PZ2Hfv7CNvt/7d+O5qNNW3jsrpnbrLPiUIU4b8xR/PyaJ1i3aiPX3X4ar09fwsrF72/d5qWnF/C3x+YC8MUj9+bsSw7nv783lQ8/+JjbrnuSD9ZsoluvTnznpi/z72dPbKmv0uJKMCeL2nvQZcCJDYXkjm7uy8vpvm9nuvXpTNt2bTj+7AG88Jd5dbbZq1cn9j1krwYHix80rA8779oureoa0LtfV95bvp7VKz+kuqqGl6ctZOBRe9fZ5qNNW7ZO79S+DUEAsHT+Gj5Yk+loe8WidbRr14Y2bXfcjrsqKlTQpzUpSotS0h1AH2CKpPuA04H2wGbgXyJiXr3tTwGuA74CDAJ+AOwELEi231CMeraU1SvW07VHx63zXbt3ZM7Ly1qwRpZPpy47s27Vxq3z61Ztos+BXbfZ7vhRB3LimQNo06aCn149ZZv1hx3Ti8XvrKZqS01R69uauUWZiIhLgBXA8cCvgS9FxKHA9cCPs7eV9FXgWuDkZNF1wPCIGATMBK5qqAxJoyXNlDRz1epVxfgaZk027dE5/Mc3HuKPd83g1PMH1lnXbZ/dOeNfh3DvrX9vodq1PKnZxsxJVRo3c3YD7pbUFwigbda6YWS6aB8REeslnUpmRLW/J39R7YDpDR00GTtjPMDgwwZH8arf/Lp068iqZeu3zq9avp4u3Tvm2MNa2rrVm+jUtcPW+U5dd2bd6o2Nbj9j2kLOv/LoT7fvsjOXjR3OhBv/xqqVHxa1rtb80rhQ8kNgWkQMIHNq3T5r3QJgV2D/ZF5kRkcbmHz6R8RFKdQxVf2GdGP5/DWsfHcdWz6pYtrv3+aoUw9o6WpZDovmrmLP7h3pstcuVLapYOjxfXjjhSV1ttkj65fdIUf05L3lHwDwuQ7tGPPjEfzprhnMn/UeOzq3KBu2G5+OR3FBvXWLgX8H/iTpLDJDSo6TtF9EzJfUAegeEf9IoZ6pqWxTybd/cTLfPeVeaqqDky44lF4H7cH/fP9pDjisG0d9pR9zZyznhrMmsmHdR0x//B/cPfYZJrxxOQBXHjeBpfNWs3nDJ5zd62dcPX4UQ0bs18LfqrzV1AQP/HI6//emkVRUiL9P+QcrFr/PqAsGsWjeat6YvoRhp/fnwEHdqK6qYdOGj5lw07MADDu9P3t068ip3ziUU79xKAC3fvcJPnz/o5b8Si2mlWVgQRRRnLNWSYvInFb3Be4GNgKPA+dHRK/sx4OSsS3uJ9Pi3Ae4iczNHIDrkm7dGzX4sMHx0ksvF+V7WHFcfMJvW7oK1kS/e3b0K9s7KuLBBw+MRx99qqBt9923y3aX11yK1qKMiF7J5Go+PbWGzM0aIuJ3wO+S6dfIXJuEzOn4kGLVy8xaVgEDh7U6fjPHzNJVejnpoDSz9BQ4Zner46A0sxQVNGZ3q+OgNLNUuUVpZpZHa3tGshAOSjNLV+nlpIPSzNJVgjnpoDSz9IjSPPXecTvFMzMrkFuUZpYeleZwtW5Rmpnl4RalmaWqBC9ROijNLE1+M8fMLL/Sy0lfozSz9AioUGGfvMeSRkqaJ2m+pGsbWH+MpFclVUk6s966akmvJ5+c/d2CW5RmlqbMg5TbfxipEhgHnAgsA2ZImhQRs7M2W0JmVIWrGzjE5ogY2MDyBjkozSxVzXTmPRSYHxELASRNBEYBW4MyIhYl67Z7bGCfeptZqmr7pMz3AbrUDkmdfEZnHaY7sDRrflmyrFDtk2O+KOn0fBu7RWlm6Sr81Ht1EcfM2ScilkvqAzwt6a2IWNDYxm5RmlmqVOAnj+VAz6z5Hnw62mteEbE8+XMh8AxwaK7tHZRmlpraTjGaYVzvGUBfSb0ltQPOAfLevSZTfidJOyXTXYCjybq22RAHpZmlqgnXKBsVEVXAFcBUYA7wh4iYJWmspNMy5WiIpGXAWcCdkmYlux8IzJT0BjANuLHe3fJt+BqlmZWkiJgMTK637Pqs6RlkTsnr7/cCcHBTynJQmll6PAqjmVkhSi8pHZRmliq3KM3M8nFQmpk1TiXazZofDzIzy8MtSjNLVSleo3SL0swsD7cozSw9Ks1xvR2UZpau0stJB6WZpasEc9JBaWYpK8FTb9/MMTPLwy1KM0tV6bUnHZRmlqLajntLjYPSzNJVejnpoDSzdJVgTjoozSxFBY4c1to4KM0sZaWXlA5KM0tV6cWkg9LMUlaCN70dlGaWptIcXcxv5piZ5eEWpZmlqgQblG5Rmpnl4xalmaWmVF9hVES0dB22m6RVwOKWrkeRdAFWt3QlrEnK9We2T0R03Z4DSHqCzN9PIVZHxMjtKa+5lEVQljNJMyNicEvXwwrnn1n58TVKM7M8HJRmZnk4KFu/8S1dAWsy/8zKjK9Rmpnl4RalmVkeDkozszwclCVM0iJJhT6TZttJ0hhJcyTd38j6CyT9Ku16WfH5zZxWQpnXFRQRNS1dF2vUZcDwiFjW0hWxdLlF2YIk9ZI0T9I9wNvAf0maIelNST/I2u4RSa9ImiVpdMvVeMcl6Q6gDzBF0nclTZf0mqQXJB3QwPanJNt0kTQimX5V0kOSdkn/G9j28F3vFiSpF7AQOAroCJwJfIvMK7GTgJsj4llJnSNiraTPATOAYyNijaRFwOCIKMfX5Vqd2r9v4BNgU0RUSRoOXBoRZ0i6IFn/FHAVcBpQCfwJOCkiNkr6LrBTRIxtie9gn41PvVve4oh4UdJPgRHAa8nyXYC+wLPAGElfTZb3TJavSb2mVms34G5JfYEA2matG0YmLEdExHpJpwL9gb8nnUG0A6anXF/bTg7Klrcx+VPATyLizuyVko4DhgNHRsQmSc8A7VOtodX3Q2BaRHw1OSt4JmvdAjKn6PsDM8n8XJ+MiHNTrqM1I1+jbD2mAhfWXr+S1F3SHmRaL+uSkOwHHNGSlTQg8zNZnkxfUG/dYuAM4B5JBwEvAkdL2g9AUgdJ+6dVUWseDspWIiL+CjwATJf0FvBHYFfgCaCNpDnAjWT+4VnLuhn4iaTXaOCsLCLmAucBD5G59nwB8KCkN8mcdvdLr6rWHHwzx8wsD7cozczycFCameXhoDQzy8NBaWaWh4PSzCwPB+UOQlK1pNclvZ28b7zzdhzrd5LOTKZ/I6l/jm2Pk3TUZyijwZ6RCukxSdKGJpb1fUlXN7WOtuNwUO44NkfEwIgYQOZd5UuyV0r6TG9pRcTFETE7xybHkXmX3axkOSh3TM8B+yWtveckTQJmS6qUdEtWD0bfgkwXcJJ+lfR09L/AHrUHkvSMpMHJ9Mikh5w3JD2VvN53CfCdpDX7JUldJT2clDFD0tHJvp+X9Nekh6TfkHn1L6dcvSpJujVZ/pSkrsmyfSU9kezzXPKmk1leftd7B5O0HE8i88YPwCBgQES8m4TNBxExRNJOZDpy+CtwKHAAmc4d9gRmAxPqHbcrcBdwTHKs2h6P7gA2RMRPk+0eAG6NiOcl7U3m1c0DgRuA5yNirKRTgIsK+DoXZveqJOnhiFgDdABmRsR3JF2fHPsKMoN+XRIR70g6HLidTCcWZjk5KHccn5P0ejL9HPBbMqfEL0fEu8nyEcAhtdcfybzT3Bc4BngwIqqBFZKebuD4RwDP1h4rItY2Uo/hQP+kJx2Ajsn77ccA/yfZ93FJ6wr4To31qlQD/D5Zfh/wp6SMo4CHssreqYAyzByUO5DNETEwe0ESGBuzFwHfjoip9bY7uRnrUQEcEREfNVCXgjWxV6VIyn2//t+BWSF8jdKyTQUuldQWQNL+kjqQ6RPz7OQa5heA4xvY90XgGEm9k307J8s/JNO5R62/At+unZFUG1zPAl9Plp0EdMpT11y9KlWQ6QSZ5JjPR8R64F1JZyVlSNIX85RhBjgora7fkLn++Kqkt4E7yZx1/Bl4J1l3Dw10PBsRq4DRZE5z3+DTU9+/AF+tvZkDjAEGJzeLZvPp3fcfkAnaWWROwZfkqWuuXpU2AkOT7zAMqO1N/DzgoqR+s4BRBfydmLn3IDOzfNyiNDPLw0FpZpaHg9LMLA8HpZlZHg5KM7M8HJRmZnk4KM3M8vj/E9ZZTzy3K8oAAAAASUVORK5CYII=\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "PNG versions of the Altair charts from the previous cell are embedded here so they will show on GitHub.\n",
        "\n",
        "<img src='https://drive.google.com/uc?id=1F7oJqpjo7d9ybofm2H_lJnngpNq7Zck3'>\n",
        "\n",
        "<img src='https://drive.google.com/uc?id=170hf5giLxTqh8bMB_11_7VH5oigp92n8'>\n"
      ],
      "metadata": {
        "id": "-Fm6LFp2NZ-7"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Features: Add punctuation"
      ],
      "metadata": {
        "id": "roZSu8cAv55M"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Using these eight punctuation marks\n",
        "punctString = '.,-?!:;\"'    # period, comma, dash, question mark, exclamation mark, colon, semi-colon, quotes\n",
        "\n",
        "# Count and add to a new column as a dictionary\n",
        "# Counts are normalized by the lengths of the articles\n",
        "articlesDF['accPunctDict'] = articlesDF.Text.apply(lambda s: {k:(v/len(s)) for k, v in Counter(s).items() if k in punctString})\n",
        "\n",
        "# Convert the new column to a dataframe\n",
        "punctDF = articlesDF.accPunctDict.apply(pd.Series)\n",
        "\n",
        "# Merge with original dataframe,\n",
        "# drop unnecessary column,\n",
        "# and fill Nans with 0 (i.e. zero counts)\n",
        "articlesDF = pd.concat([articlesDF, punctDF], axis=1)\n",
        "articlesDF = articlesDF.drop(['accPunctDict'], axis=1)\n",
        "articlesDF = articlesDF.fillna(0)\n",
        "\n",
        "# Rename the columns to something readable\n",
        "articlesDF = articlesDF.rename(columns={'\"': 'DblQuote', '.': 'Period', '-': 'Dash', ',': 'Comma',\n",
        "                                            '!': 'ExclMark', ';': 'SemiCol', ':': 'Colon', '?': 'QuesMark'})\n",
        "\n",
        "articlesDF.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 250
        },
        "id": "u0KGsUvywByS",
        "outputId": "931b19ca-10c9-4ee1-8240-2b94f3bfefc3"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                Text Domain  intLabel  \\\n",
              "0  Supreme Court Forces Schools Into Socialism\\n\\...    edu         1   \n",
              "1  Brexit fist lost\\nThe oil-rich Gulf country Qa...    biz         1   \n",
              "2  Leftist universities seek to put women in posi...    edu         1   \n",
              "3  New intelligence suggests ISIS rigging laptop ...   tech         1   \n",
              "4  Microsoft Aims to spread liberalism on the suc...    edu         1   \n",
              "\n",
              "      Comma    Period  DblQuote      Dash  QuesMark     Colon  ExclMark  \\\n",
              "0  0.004603  0.011507  0.013809  0.001151  0.000000  0.000000       0.0   \n",
              "1  0.001245  0.008717  0.000000  0.001245  0.000000  0.000000       0.0   \n",
              "2  0.007067  0.005300  0.003534  0.001767  0.001767  0.000000       0.0   \n",
              "3  0.003831  0.006386  0.002554  0.000000  0.000000  0.000000       0.0   \n",
              "4  0.001346  0.006729  0.013459  0.001346  0.000000  0.001346       0.0   \n",
              "\n",
              "   SemiCol  \n",
              "0      0.0  \n",
              "1      0.0  \n",
              "2      0.0  \n",
              "3      0.0  \n",
              "4      0.0  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-963127e4-637e-431c-b376-b10034ce5bb1\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>Text</th>\n",
              "      <th>Domain</th>\n",
              "      <th>intLabel</th>\n",
              "      <th>Comma</th>\n",
              "      <th>Period</th>\n",
              "      <th>DblQuote</th>\n",
              "      <th>Dash</th>\n",
              "      <th>QuesMark</th>\n",
              "      <th>Colon</th>\n",
              "      <th>ExclMark</th>\n",
              "      <th>SemiCol</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Supreme Court Forces Schools Into Socialism\\n\\...</td>\n",
              "      <td>edu</td>\n",
              "      <td>1</td>\n",
              "      <td>0.004603</td>\n",
              "      <td>0.011507</td>\n",
              "      <td>0.013809</td>\n",
              "      <td>0.001151</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Brexit fist lost\\nThe oil-rich Gulf country Qa...</td>\n",
              "      <td>biz</td>\n",
              "      <td>1</td>\n",
              "      <td>0.001245</td>\n",
              "      <td>0.008717</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.001245</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>Leftist universities seek to put women in posi...</td>\n",
              "      <td>edu</td>\n",
              "      <td>1</td>\n",
              "      <td>0.007067</td>\n",
              "      <td>0.005300</td>\n",
              "      <td>0.003534</td>\n",
              "      <td>0.001767</td>\n",
              "      <td>0.001767</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>New intelligence suggests ISIS rigging laptop ...</td>\n",
              "      <td>tech</td>\n",
              "      <td>1</td>\n",
              "      <td>0.003831</td>\n",
              "      <td>0.006386</td>\n",
              "      <td>0.002554</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Microsoft Aims to spread liberalism on the suc...</td>\n",
              "      <td>edu</td>\n",
              "      <td>1</td>\n",
              "      <td>0.001346</td>\n",
              "      <td>0.006729</td>\n",
              "      <td>0.013459</td>\n",
              "      <td>0.001346</td>\n",
              "      <td>0.000000</td>\n",
              "      <td>0.001346</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-963127e4-637e-431c-b376-b10034ce5bb1')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-963127e4-637e-431c-b376-b10034ce5bb1 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-963127e4-637e-431c-b376-b10034ce5bb1');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# NOTE: The Altair chart 'predChart' will not show on GitHub as\n",
        "# explained in the intro and the appendix. It is embedded in the\n",
        "# following Markdown cell as a workaround. The second chart is\n",
        "# generated by Matplotlib and isn't affected.\n",
        "\n",
        "\n",
        "trainChart, predChart, confMtxPlot = modelData(articlesDF, stratify=False)\n",
        "# Not displaying the same training data split\n",
        "# Just the weights chart\n",
        "#predChart.display()\n",
        "\n",
        "# Show the confusion matrix\n",
        "confMtxPlot;"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 447
        },
        "id": "BfAdmnO-xn3O",
        "outputId": "51288922-866e-4ce0-fcca-e62562d37eee"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training and test dataframe shapes: (784, 11) (196, 11)\n",
            "Fake (1) / real (0) label counts in the training data: Counter({1: 405, 0: 379})\n",
            "\n",
            "\n",
            "Random f1 score: 0.5340909090909091 ... Accuracy: 0.5816326530612245\n",
            "Logistic regression f1 score: 0.700507614213198 ... Accuracy: 0.6989795918367347\n",
            "SVC f1 score: 0.7076923076923078 ... Accuracy: 0.7091836734693877\n",
            "\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "A PNG version of the Altair chart from the previous cell is embedded here so it will show on GitHub.\n",
        "\n",
        "<img src='https://drive.google.com/uc?id=1zBygdDVKBBeZ_nRiPDGE1M4CEhirvY6t'>\n"
      ],
      "metadata": {
        "id": "Sg6CLV0DNk9Z"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Features: Add Empath categories"
      ],
      "metadata": {
        "id": "ytuNePTe2Y2_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Create six new categories\n",
        "\n",
        "# 1- Hedge ==> false\n",
        "lexicon.create_category(\"hedge\",['wonder', 'could', 'sort_of', 'maybe', 'possibly', 'just',\n",
        "                                 'feel', 'reckon', 'generally_agreed'])\n",
        "# 2- Modality ==> true/false\n",
        "lexicon.create_category(\"modality\",[\"could\",\"may\",\"possibly\"])\n",
        "# 3- Negation ==> truthful\n",
        "lexicon.create_category(\"negation\",['no', 'not', 'never', 'none', 'nobody', \"n't\", 'hardly', 'rarely', 'seldom', 'scarcely'])\n",
        "# 4- Polarity ==> truthful\n",
        "lexicon.create_category(\"polarity\", ['accurate', 'inaccurate', 'uncertain', 'certain', 'likely', 'unlikely',\n",
        "                                                              'probable', 'improbable', 'possible', 'impossible'])\n",
        "# 5- Spatial ==> truthful\n",
        "lexicon.create_category(\"spatial\",[\"big\",\"small\",\"circular\"])\n",
        "# 6- Subjectivity ==> truthful\n",
        "lexicon.create_category(\"subjectivity\", ['abhorrent', 'abominable', 'brutal', 'coercive', 'disconcerting',\n",
        "                                            'grateful', 'indecision', 'miserable', 'plausible', 'reflective'])\n",
        "\n",
        "\n",
        "# List of categories (total of 14) to use as features, including six new ones\n",
        "featCategs = ['anticipation', 'hedge', 'joy', 'modality', 'negation', 'negative_emotion', 'noise', 'party', 'polarity',\n",
        "                          'positive_emotion', 'spatial', 'subjectivity', 'torment', 'vehicle']\n",
        "\n"
      ],
      "metadata": {
        "id": "LfCpz5r92PpX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fe914227-b651-467f-c921-b7ed9221953e"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[\"feel\", \"though\", \"wonder\", \"maybe\", \"thought\", \"wish\", \"probably\", \"Because\", \"might\", \"Maybe\", \"actually\", \"knew\", \"honestly\", \"knowing\", \"mean\", \"should\", \"wanted\", \"seem\", \"anything\", \"Sometimes\", \"guess\", \"though\", \"Or\", \"possibly\", \"suppose\", \"imagine\", \"anymore\", \"anyway\", \"If\", \"always\", \"sure\", \"bet\", \"thinking\", \"someone\", \"understand\", \"realize\", \"cause\", \"else\", \"sometimes\", \"Honestly\", \"care\", \"better\", \"anyone\", \"doubt\", \"realise\", \"meant\", \"either\", \"Even\", \"help\", \"try\", \"does\", \"better\", \"Now\", \"only\", \"cause\", \"figured\", \"right\", \"feels\"]\n",
            "[\"might\", \"may\", \"possibly\", \"should\", \"'d\", \"must\", \"able\", \"wanted\", \"ought\", \"willing\", \"surely\", \"couldn\\u2019t\", \"otherwise\", \"wish\", \"wo\", \"Could\", \"If\", \"wouldn\\u2019t\", \"other_ways\", \"suppose\", \"potentially\", \"capability\", \"Though\", \"wished\", \"couldn\\u2019t\", \"Nor\", \"though\", \"needed\", \"enough\", \"ability\", \"didn\\u2019t\", \"couldn&acirc;&euro;&trade;t\", \"potential\", \"Although\", \"meant\", \"although\", \"any\", \"even_I\", \"couldn&rsquo;t\", \"easiest_way\", \"wanting\", \"Perhaps\", \"Yet\", \"inability\", \"maybe\", \"does\", \"certain_people\", \"however\", \"Or\", \"refuse\", \"rightly\", \"Should\", \"Did\", \"can&acirc;&euro;&trade;t\", \"can\\u2019t\", \"perhaps\", \"personally\", \"tend\", \"potential\", \"allow\", \"enough_people\", \"wouldn\\u2019t\", \"knew\", \"Sometimes\", \"certainly\", \"Would\", \"though\", \"However\", \"supposed\", \"Most_people\", \"expect\", \"hoped\", \"therefore\", \"able\", \"intend\", \"expected\", \"either\", \"impossible\", \"yet\", \"surly\", \"doubted\"]\n",
            "[\"nobody\", \"none\", \"Nobody\", \"anybody\", \"one\", \"nothing\", \"anyone\", \"None\", \"No\", \"anything\", \"else\", \"anyway\", \"noone\", \"only_one_person\", \"someone\", \"either\", \"either\", \"unless\", \"somebody\", \"actually\", \"very_few_people\", \"Not\", \"whatever\", \"besides\", \"everybody\", \"ether\", \"Because\", \"any\", \"cause\", \"everyone\", \"anyways\", \"only\", \"person\", \"Nothing\", \"always\", \"though\", \"cause\", \"too_many_people\", \"probably\", \"nt\", \"hardly\", \"anymore\", \"neither\", \"other_person\", \"honestly\", \"Obviously\", \"many_people\", \"mean\", \"whatever\", \"Neither\", \"just_one_person\", \"guess\", \"anyways\", \"If\", \"reason\", \"people\", \"only\", \"almost_no_one\", \"care\", \"thing\", \"obviously\", \"only_thing\", \"only_one\", \"yet\", \"why\", \"otherwise\", \"only_person\", \"single_soul\", \"course\", \"cause\", \"no\", \"hardly_anyone\", \"other_people\"]\n",
            "[\"unlikely\", \"certain\", \"logical\", \"probable\", \"accurate\", \"impossible\", \"plausible\", \"likely\", \"vital\", \"certainly\", \"effective\", \"critical\", \"difficult\", \"unfortunate\", \"beneficial\", \"crucial\", \"dangerous\", \"reasonable\", \"necessary\", \"rational\", \"disastrous\", \"practical\", \"absurd\", \"doubted\", \"accurately\", \"adequate\", \"Though\", \"although\", \"imperative\", \"successful\", \"remarkable\", \"illogical\", \"significant\", \"troubling\", \"definite\", \"satisfactory\", \"improbable\", \"unpleasant\", \"preferable\", \"perceived\", \"however\", \"useful\", \"reliable\", \"uncommon\", \"risky\", \"Perhaps\", \"surely\", \"possibility\", \"compromised\", \"ideal\", \"intentional\", \"doubtful\", \"unwise\", \"However\", \"bizarre\", \"ludicrous\", \"inclined\", \"perhaps\", \"otherwise\", \"possibly\", \"entirely\", \"definitely\", \"questionable\", \"miraculous\", \"mundane\", \"complex\", \"wise\", \"undoubtedly\", \"problematic\", \"promising\", \"Although\", \"thorough\", \"convinced\", \"desirable\", \"inaccurate\", \"keen\", \"frightening\", \"convenient\", \"feasible\", \"inconvenient\", \"suspected\", \"Furthermore\", \"vague\", \"necessity\", \"speculation\", \"tricky\", \"conclude\", \"unavoidable\", \"initially\", \"fortunate\", \"adamant\", \"troublesome\", \"discreet\", \"prudent\", \"Though\", \"harmful\", \"costly\", \"laughable\", \"outrageous\", \"efficient\"]\n",
            "[\"small\", \"big\", \"large\", \"huge\", \"tiny\", \"massive\", \"gigantic\", \"rectangular\", \"giant\", \"enormous\", \"circular\", \"size\", \"compact\", \"bulky\", \"circular\", \"wooden\", \"perfect_size\", \"shiny\", \"ginormous\", \"Small\", \"average_size\", \"humongous\", \"oval\", \"smaller\", \"roomy\", \"bigger\", \"sturdy\", \"thin\", \"rectangle\", \"spacious\", \"rustic\", \"medium\", \"shabby\", \"dainty\", \"shaped\", \"small_one\", \"medium\", \"broad\", \"plush\", \"sized\", \"modern\", \"round\", \"tall\", \"cramped\", \"luxurious\", \"big_one\", \"sleek\", \"quaint\", \"dingy\", \"white\", \"rounded\", \"miniature\", \"hairy\", \"square\", \"furry\", \"colorful\", \"dusty\", \"same_size\", \"HUGE\", \"small\", \"larger\", \"antique\", \"marble\", \"oak\", \"decorative\", \"smallest\", \"ornate\", \"light\", \"solid\", \"walnut\", \"decent_size\", \"medium_size\", \"sizes\", \"exterior\", \"structure\", \"shape\", \"elaborate\", \"build\", \"narrow\", \"flimsy\", \"golden\", \"fashioned\", \"decoration\", \"minuscule\", \"dark\", \"slim\", \"thick\", \"bronze\", \"slender\", \"crooked\", \"good_size\", \"plain\", \"Huge\", \"elegant\", \"silver\", \"colourful\", \"colored\", \"white_stone\", \"symmetrical\"]\n",
            "[\"devastating\", \"unhappy\", \"grateful\", \"brutal\", \"miserable\", \"inadequate\", \"intolerable\", \"unreasonable\", \"unfortunate\", \"rational\", \"unpleasant\", \"desirable\", \"critical\", \"tolerant\", \"unavoidable\", \"dreadful\", \"appalling\", \"meaningless\", \"humiliating\", \"dependent\", \"problematic\", \"remarkable\", \"laughable\", \"logical\", \"troubling\", \"frightening\", \"admirable\", \"insignificant\", \"shameful\", \"illogical\", \"inconvenient\", \"troublesome\", \"morbid\", \"inconsequential\", \"rewarding\", \"enlightened\", \"tolerable\", \"wretched\", \"disastrous\", \"distasteful\", \"pessimistic\", \"terrifying\", \"disconcerting\", \"questionable\", \"barbaric\", \"absurd\", \"unsettling\", \"isolated\", \"unimportant\", \"preferable\", \"satisfying\", \"apathetic\", \"tedious\", \"cynical\", \"probable\", \"disappointing\", \"infuriating\", \"predictable\", \"terrifying\", \"bizarre\", \"agreeable\", \"unpredictable\", \"optimistic\", \"ludicrous\", \"thankful\", \"resentful\", \"flawed\", \"plausible\", \"daunting\", \"discouraged\", \"cruel\", \"gruesome\", \"callous\", \"merciful\", \"certain\", \"worrying\", \"complacent\", \"cooperative\", \"trivial\", \"worrisome\", \"alarming\", \"vulnerable\", \"preposterous\", \"affected\", \"despised\", \"foolish\", \"incompetent\", \"beneficial\", \"terrible\", \"impractical\", \"profound\", \"fortunate\", \"irrational\", \"practical\", \"costly\", \"bothersome\", \"unnerving\", \"appreciative\", \"devoted\", \"mundane\"]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Count and add to a new column as a dictionary\n",
        "# Counts are normalized by the lengths of the articles\n",
        "articlesDF['lexDict'] = articlesDF.Text.apply(lambda s: lexicon.analyze(s, categories=featCategs, normalize=True))\n",
        "\n",
        "# Convert the new column to a dataframe\n",
        "lexDF = articlesDF.lexDict.apply(pd.Series)\n",
        "\n",
        "# Merge with original dataframe,\n",
        "# drop unnecessary column,\n",
        "# and fill Nans with 0 (i.e. zero counts)\n",
        "articlesDF = pd.concat([articlesDF, lexDF], axis=1)\n",
        "articlesDF = articlesDF.drop(['lexDict'], axis=1)\n",
        "articlesDF = articlesDF.fillna(0)\n",
        "\n",
        "articlesDF.head(1)"
      ],
      "metadata": {
        "id": "CwruMmBVjdbz",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 226
        },
        "outputId": "256e1c7f-75bd-4274-f567-a5fba30571e2"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                                                Text Domain  intLabel  \\\n",
              "0  Supreme Court Forces Schools Into Socialism\\n\\...    edu         1   \n",
              "\n",
              "      Comma    Period  DblQuote      Dash  QuesMark  Colon  ExclMark  ...  \\\n",
              "0  0.004603  0.011507  0.013809  0.001151       0.0    0.0       0.0  ...   \n",
              "\n",
              "   negation  negative_emotion  noise  party  polarity  positive_emotion  \\\n",
              "0  0.007752               0.0    0.0    0.0  0.015504               0.0   \n",
              "\n",
              "   spatial  subjectivity  torment  vehicle  \n",
              "0      0.0      0.007752      0.0      0.0  \n",
              "\n",
              "[1 rows x 25 columns]"
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-60618dc5-fb93-44d2-8a8e-648a10ba22d0\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>Text</th>\n",
              "      <th>Domain</th>\n",
              "      <th>intLabel</th>\n",
              "      <th>Comma</th>\n",
              "      <th>Period</th>\n",
              "      <th>DblQuote</th>\n",
              "      <th>Dash</th>\n",
              "      <th>QuesMark</th>\n",
              "      <th>Colon</th>\n",
              "      <th>ExclMark</th>\n",
              "      <th>...</th>\n",
              "      <th>negation</th>\n",
              "      <th>negative_emotion</th>\n",
              "      <th>noise</th>\n",
              "      <th>party</th>\n",
              "      <th>polarity</th>\n",
              "      <th>positive_emotion</th>\n",
              "      <th>spatial</th>\n",
              "      <th>subjectivity</th>\n",
              "      <th>torment</th>\n",
              "      <th>vehicle</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>Supreme Court Forces Schools Into Socialism\\n\\...</td>\n",
              "      <td>edu</td>\n",
              "      <td>1</td>\n",
              "      <td>0.004603</td>\n",
              "      <td>0.011507</td>\n",
              "      <td>0.013809</td>\n",
              "      <td>0.001151</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>...</td>\n",
              "      <td>0.007752</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.015504</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.007752</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>1 rows × 25 columns</p>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-60618dc5-fb93-44d2-8a8e-648a10ba22d0')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-60618dc5-fb93-44d2-8a8e-648a10ba22d0 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-60618dc5-fb93-44d2-8a8e-648a10ba22d0');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# NOTE: The Altair chart 'predChart' will not show on GitHub as\n",
        "# explained in the intro and the appendix. It is embedded in the\n",
        "# following Markdown cell as a workaround. The second chart is\n",
        "# generated by Matplotlib and isn't affected.\n",
        "\n",
        "trainChart, predChart, confMtxPlot = modelData(articlesDF, stratify=False)\n",
        "# Not displaying the same training data split\n",
        "# Just the weights chart\n",
        "#predChart.display()\n",
        "\n",
        "# Show the confusion matrix\n",
        "confMtxPlot;"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 444
        },
        "id": "3MvJtvTHyA96",
        "outputId": "2e6d5e44-b57d-4643-b27d-780fac3ee4aa"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training and test dataframe shapes: (784, 25) (196, 25)\n",
            "Fake (1) / real (0) label counts in the training data: Counter({1: 405, 0: 379})\n",
            "\n",
            "\n",
            "Random f1 score: 0.5340909090909091 ... Accuracy: 0.5816326530612245\n",
            "Logistic regression f1 score: 0.6979166666666667 ... Accuracy: 0.7040816326530612\n",
            "SVC f1 score: 0.7046632124352332 ... Accuracy: 0.7091836734693877\n",
            "\n",
            "\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "A PNG version of the Altair chart from the previous cell is embedded here so it will show on GitHub.\n",
        "\n",
        "<img src='https://drive.google.com/uc?id=1p75OAdBKfopXUxmbzk2mzTEqzWCuk-GF'>"
      ],
      "metadata": {
        "id": "WLFaHrEkbmi0"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### **Scattertext**"
      ],
      "metadata": {
        "id": "ju5a3z36j4JZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Form a complete list of terms from the Empath categories used above\n",
        "termList = []\n",
        "for eaCateg in featCategs:\n",
        "  termList.extend(lexicon.cats[eaCateg][:10])\n",
        "# Add terms from feature weights above\n",
        "moreTerms = ['said', 'source', 'insider', 'year', 'night', 'day', 'revealed',\n",
        "             'like', 'people', 'world', 'old', 'way', 'wants', 'like']\n",
        "termList.extend(moreTerms)\n",
        "# Convert to a set to remove possible duplicates\n",
        "termList = set(termList)\n",
        "\n",
        "\n",
        "# Read in the data again to get the original labels\n",
        "scatterDF = pd.read_csv(inFile, index_col=0)\n",
        "\n",
        "scatterDF = scatterDF.assign(\n",
        "    parse=lambda scatterDF: scatterDF.Text.apply(st.whitespace_nlp_with_sentences)\n",
        ")\n",
        "\n",
        "corpus = st.CorpusFromParsedDocuments(\n",
        "    scatterDF, category_col='Label', parsed_col='parse'\n",
        ").build().get_unigram_corpus().compact(st.AssociationCompactor(2000))\n",
        "\n",
        "# Generate the scatter plot data\n",
        "scatterData = st.produce_scattertext_explorer(\n",
        "    corpus,\n",
        "    category='fake', category_name='Fake', not_category_name='Real',\n",
        "    minimum_term_frequency=0, pmi_threshold_coefficient=0,\n",
        "    return_data=True,\n",
        "    transform=st.Scalers.dense_rank\n",
        ")\n",
        "# Put it into a dataframe, filter out labels\n",
        "# for terms not in termList\n",
        "scatterPlotDF = pd.DataFrame(scatterData['data'])\n",
        "scatterPlotDF['plotTerm'] = scatterPlotDF['term'].apply(lambda x: x if x in termList else ' ')\n",
        "# Also filter out labels for terms too close\n",
        "# to the origin and so not legible\n",
        "scatterPlotDF['plotTerm'] = scatterPlotDF[['ox','oy', 'plotTerm']] \\\n",
        "                                .apply(lambda x: x['plotTerm'] if (x['oy']>0.25) else ' ', axis=1)\n",
        "\n",
        "scatterPlot = alt.Chart(scatterPlotDF).mark_circle().encode(\n",
        "    x = alt.X('x:Q', title='\\u21e6 less with real | more with real \\u21e8', axis=alt.Axis(tickMinStep=0.1)),\n",
        "    y = alt.Y('y:Q', title='\\u21e6 less with fake | more with fake \\u21e8', axis=alt.Axis(tickMinStep=0.1)),\n",
        "    color = alt.Color('s', scale=alt.Scale(scheme='redyellowgreen', reverse=True),\n",
        "                      legend=alt.Legend(title=[\"Fake vs real\", \"frequency\"])),\n",
        ").properties(\n",
        "    width = 500,\n",
        "    height = 500\n",
        ")\n",
        "\n",
        "text = alt.Chart(scatterPlotDF).mark_text(\n",
        "    align='left',\n",
        "    baseline='middle',\n",
        "    dx=7,\n",
        "    color='black'\n",
        ").encode(\n",
        "    x = alt.X('x:Q'),\n",
        "    y = alt.Y('y:Q'),\n",
        "    text='plotTerm'\n",
        ").properties(\n",
        "    title = 'How frequently are terms used with each type of news?',\n",
        "    width = 500,\n",
        "    height = 500\n",
        ")\n",
        "\n",
        "# NOTE: This Altair chart will not show on GitHub as explained\n",
        "# in the intro and the appendix. It is embedded in the following\n",
        "# Markdown cell as a workaround.\n",
        "\n",
        "#(scatterPlot + text).configure_axis(\n",
        "#    labelFontSize=10,\n",
        "#    titleFontSize=12\n",
        "#).configure_title(\n",
        "#    fontSize=15,\n",
        "#    anchor='start'\n",
        "#).configure_legend(\n",
        "#    titleFontSize=12\n",
        "#)\n"
      ],
      "metadata": {
        "id": "r4QP_X-AjyrP"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "A PNG version of the Altair chart from the previous cell is embedded here so it will show on GitHub.\n",
        "\n",
        "<img src='https://drive.google.com/uc?id=1HBVkgu_Gs6Dpxyqiyfy9lLJ72YgICdsH'>\n"
      ],
      "metadata": {
        "id": "mBJfOHl_cfSp"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "###Appendix: Altair images\n",
        "\n",
        "Altair-generated images don't show on GitHub. The reason is explained here: https://stackoverflow.com/questions/71346406/why-are-my-altair-data-visualizations-not-showing-up-in-github\n",
        "\n",
        "So this notebook contains a workaround for several Altair charts above, embedding PNG versions of them into Markdown cells. The method is explained here: https://medium.com/analytics-vidhya/embedding-your-image-in-google-colab-markdown-3998d5ac2684"
      ],
      "metadata": {
        "id": "-euFYJdALHkr"
      }
    }
  ]
}