German_and_Adult.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "German_and_Adult.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "P-ygJNn1jeCW",
        "HkPZn81Yjxp3",
        "rh6APOPSkEh0",
        "7YskGtCRkILJ",
        "869-ZIWmkT6q",
        "29ovdpaHmc23",
        "Z0J5MeZ5mym5",
        "nBFBQNVKnApI",
        "xPfopYgWnIyb",
        "X8R1JmQInV51",
        "d3j_ZWBAryu1"
      ]
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "#934G5: Machine Learning\n",
        "# CandNo: 244106\n"
      ],
      "metadata": {
        "id": "495QWDhjU10a"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install 'aif360[LFR]'"
      ],
      "metadata": {
        "id": "vGaJ7tjQU3Wu"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "cd /usr/local/lib/python3.7/dist-packages/aif360/data/raw/adult"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dNOel_8viQFw",
        "outputId": "41313eec-4f4b-4c81-98d8-3d2379544371"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/aif360/data/raw/adult\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!wget https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data\n",
        "!wget https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.names\n",
        "!wget https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.test"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pX4R_V7HiRoS",
        "outputId": "092d201e-4a81-4eba-ab67-fc96e631c1a5"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2022-05-13 11:26:54--  https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data\n",
            "Resolving archive.ics.uci.edu (archive.ics.uci.edu)... 128.195.10.252\n",
            "Connecting to archive.ics.uci.edu (archive.ics.uci.edu)|128.195.10.252|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 3974305 (3.8M) [application/x-httpd-php]\n",
            "Saving to: ‘adult.data.3’\n",
            "\n",
            "adult.data.3        100%[===================>]   3.79M  3.13MB/s    in 1.2s    \n",
            "\n",
            "2022-05-13 11:26:56 (3.13 MB/s) - ‘adult.data.3’ saved [3974305/3974305]\n",
            "\n",
            "--2022-05-13 11:26:56--  https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.names\n",
            "Resolving archive.ics.uci.edu (archive.ics.uci.edu)... 128.195.10.252\n",
            "Connecting to archive.ics.uci.edu (archive.ics.uci.edu)|128.195.10.252|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 5229 (5.1K) [application/x-httpd-php]\n",
            "Saving to: ‘adult.names.3’\n",
            "\n",
            "adult.names.3       100%[===================>]   5.11K  --.-KB/s    in 0s      \n",
            "\n",
            "2022-05-13 11:26:57 (230 MB/s) - ‘adult.names.3’ saved [5229/5229]\n",
            "\n",
            "--2022-05-13 11:26:57--  https://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.test\n",
            "Resolving archive.ics.uci.edu (archive.ics.uci.edu)... 128.195.10.252\n",
            "Connecting to archive.ics.uci.edu (archive.ics.uci.edu)|128.195.10.252|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 2003153 (1.9M) [application/x-httpd-php]\n",
            "Saving to: ‘adult.test.3’\n",
            "\n",
            "adult.test.3        100%[===================>]   1.91M  1.81MB/s    in 1.1s    \n",
            "\n",
            "2022-05-13 11:26:59 (1.81 MB/s) - ‘adult.test.3’ saved [2003153/2003153]\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "cd /usr/local/lib/python3.7/dist-packages/aif360/data/raw/german"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "64vDFy44iXd5",
        "outputId": "24e89afc-9f4b-4dbd-9dba-03f31bb353df"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/aif360/data/raw/german\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!wget https://archive.ics.uci.edu/ml/machine-learning-databases/statlog/german/german.data\n",
        "!wget https://archive.ics.uci.edu/ml/machine-learning-databases/statlog/german/german.doc"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HIfBMeQBia-7",
        "outputId": "6c769eeb-ee0d-4865-e858-64913c583471"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2022-05-13 11:26:59--  https://archive.ics.uci.edu/ml/machine-learning-databases/statlog/german/german.data\n",
            "Resolving archive.ics.uci.edu (archive.ics.uci.edu)... 128.195.10.252\n",
            "Connecting to archive.ics.uci.edu (archive.ics.uci.edu)|128.195.10.252|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 79793 (78K) [application/x-httpd-php]\n",
            "Saving to: ‘german.data.3’\n",
            "\n",
            "german.data.3       100%[===================>]  77.92K   260KB/s    in 0.3s    \n",
            "\n",
            "2022-05-13 11:27:00 (260 KB/s) - ‘german.data.3’ saved [79793/79793]\n",
            "\n",
            "--2022-05-13 11:27:00--  https://archive.ics.uci.edu/ml/machine-learning-databases/statlog/german/german.doc\n",
            "Resolving archive.ics.uci.edu (archive.ics.uci.edu)... 128.195.10.252\n",
            "Connecting to archive.ics.uci.edu (archive.ics.uci.edu)|128.195.10.252|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 4679 (4.6K) [application/x-httpd-php]\n",
            "Saving to: ‘german.doc.3’\n",
            "\n",
            "german.doc.3        100%[===================>]   4.57K  --.-KB/s    in 0s      \n",
            "\n",
            "2022-05-13 11:27:00 (231 MB/s) - ‘german.doc.3’ saved [4679/4679]\n",
            "\n"
          ]
        }
      ]
    },

    {
      "cell_type": "code",
      "source": [
        "# NOTE: import (run cell) twice\n",
        "\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# AIF360 imports\n",
        "from aif360.datasets import AdultDataset\n",
        "from aif360.algorithms.preprocessing.optim_preproc_helpers.data_preproc_functions import load_preproc_data_adult\n",
        "from aif360.algorithms.preprocessing.optim_preproc_helpers.data_preproc_functions import load_preproc_data_german\n",
        "\n",
        "from aif360.metrics import ClassificationMetric\n",
        "from aif360.algorithms.preprocessing.reweighing import Reweighing\n",
        "\n",
        "# Scaler for normalisation\n",
        "from sklearn.preprocessing import StandardScaler  \n",
        "from sklearn.model_selection import *\n",
        "\n",
        "# models\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "\n",
        "np.random.seed(0) "
      ],
      "metadata": {
        "id": "0v-LcV6VidUZ"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Adult\n",
        "\n",
        "##Step 1. Prepare the dataset"
      ],
      "metadata": {
        "id": "HHXArMPuihT9"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# label the sensitive feature to be investigated\n",
        "privileged_groups = [{'sex': 1}]\n",
        "unprivileged_groups = [{'sex': 0}]\n",
        "\n",
        "# load in the dataset\n",
        "dataset_orig = load_preproc_data_adult(['sex'])\n",
        "\n",
        "# split training and test data\n",
        "train, test = dataset_orig.split([0.7], shuffle=True)\n",
        "\n",
        "# normalise dataset\n",
        "ss = StandardScaler()\n",
        "# try actually normalising the whole 'train' and 'test' dataset objects\n",
        "train.features = ss.fit_transform(train.features)\n",
        "X_train = train.features\n",
        "y_train = train.labels.ravel()\n",
        "test.features = ss.transform(test.features) \n",
        "X_test = test.features\n",
        "y_test = test.labels.ravel()\n",
        "\n",
        "# Viktoriia's method: does not actually change the train or test objects...\n",
        "#X_train = ss.fit_transform(train.features)\n",
        "#y_train = train.labels.ravel()\n",
        "#X_test = ss.transform(test.features) \n",
        "#y_test = test.labels.ravel()"
      ],
      "metadata": {
        "id": "8zm_NMPfipfB"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# functions to get folds and to perform 5-fold cross-validation\n",
        "\n",
        "def get_folds(X):\n",
        "\n",
        "  # split into 5 folds \n",
        "  kf = KFold(n_splits=5)\n",
        "  kf.get_n_splits(X)\n",
        "\n",
        "  # return the indices of all 5 splits\n",
        "  fold_indices = kf.split(X)\n",
        "\n",
        "  # iterate through and put into dictionary\n",
        "  folds = {}\n",
        "  for i, (train_index, test_index) in enumerate(fold_indices):\n",
        "    folds[i] = (train_index, test_index)\n",
        "\n",
        "  return folds\n",
        "\n",
        "def perform_cross_fold_accuracy(folds, X, y, reg_strength, train_weights=None, reweighed = False):\n",
        "  \"\"\"\n",
        "    only pass in X_train and y_train into this function as we only want to \n",
        "  perform cross validation on the training set.\n",
        "  \"\"\"\n",
        "  accuracies = []\n",
        "\n",
        "  # loop through dictionary and get accuracy score for each fold\n",
        "  for fold_num, indices in folds.items():\n",
        "    X_train, X_test = X[indices[0]], X[indices[1]]\n",
        "    y_train, y_test = y[indices[0]], y[indices[1]]\n",
        "\n",
        "    # if train_weights not equal to None, then get relevant subset\n",
        "    if reweighed == False:\n",
        "      sample_weights = None\n",
        "    else:\n",
        "      sample_weights = train_weights[indices[0]]\n",
        "\n",
        "    # define and fit classifier on the training data\n",
        "    clf = LogisticRegression(C=reg_strength, solver='liblinear', random_state=1)  \n",
        "    clf.fit(X_train,y_train,sample_weight=sample_weights)\n",
        "\n",
        "    # get predictions \n",
        "    pred = clf.predict(X_test)\n",
        "\n",
        "    # compute the accuracy and append to list\n",
        "    accuracy = clf.score(X_test, y_test)\n",
        "    accuracies.append(accuracy)\n",
        "\n",
        "  # return the mean and standard deviation of accuracies for all 5 folds\n",
        "  return np.mean(accuracies), np.std(accuracies)"
      ],
      "metadata": {
        "id": "9I6IZPgoiyhF"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# create folds dict\n",
        "folds = get_folds(X_train)"
      ],
      "metadata": {
        "id": "z2c4dkk2i1hX"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "###Step 2. C parameter sweep for accuracy"
      ],
      "metadata": {
        "id": "D7_cKIB4jMFw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def parameter_sweep_accuracy(c_s, reweighed=False):\n",
        "\n",
        "  accs = []\n",
        "  stds = []\n",
        "\n",
        "  # loop through C values and perform 5-fold cross validation on training dataset\n",
        "  for c_value in c_s:\n",
        "    if reweighed:\n",
        "      acc, std = perform_cross_fold_accuracy(folds, X_train, y_train, reg_strength=c_value, train_weights=train.instance_weights, reweighed=True)\n",
        "    else:\n",
        "      acc, std = perform_cross_fold_accuracy(folds, X_train, y_train, reg_strength=c_value)\n",
        "\n",
        "    accs.append(acc)\n",
        "    stds.append(std)\n",
        "\n",
        "  return accs, stds"
      ],
      "metadata": {
        "id": "0f-ARqQZi3ul"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# provide a list of possible regularisation values\n",
        "c_values = np.logspace(-15, 15, num=50)\n",
        "#c_values = np.logspace(-6, 1, num=50)\n",
        "\n",
        "# perform hyperparameter sweep over reg strengths\n",
        "accs1, stds1 = parameter_sweep_accuracy(c_values)"
      ],
      "metadata": {
        "id": "EYeNgsf4i53Z"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# graph of regularisation strength versus accuracy\n",
        "\n",
        "def plot_graph(c_s, metric, standard_deviations, metric_name, dataset):\n",
        "  # plot accuracy against log of regularisation strength\n",
        "  plt.plot(np.log10(c_s), metric)\n",
        "  plt.errorbar(np.log10(c_s), metric, yerr = standard_deviations, fmt='.', ecolor = 'cyan', color='black')\n",
        "  plt.title(f'Effect of regularisation strength on {metric_name}: {dataset}')\n",
        "  # C is the inverse of regularization strength\n",
        "  plt.xlabel(\" $Log_{10}$(C)\")\n",
        "  if metric_name == 'accuracy':\n",
        "    plt.ylabel(\"Accuracy\")\n",
        "  if metric_name == 'fairness':\n",
        "    plt.ylabel(\"Equality of Opportunity\")\n",
        "  if metric_name == 'both':\n",
        "    plt.ylabel('Z')    \n",
        "  plt.show()"
      ],
      "metadata": {
        "id": "JB3JCuoPi7z2"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plot_graph(c_values, accs1, stds1, 'accuracy', 'Adult')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "oOVQFrBIjDpH",
        "outputId": "dd92d648-0487-4af4-e4d0-35605b430420"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# find the most accurate\n",
        "def find_accuracy_c_value(accuracies, c_s):\n",
        "  print(f'The maximum accuracy achieved during 5-fold cross-val is {max(accuracies)}')\n",
        "  index = accuracies.index(max(accuracies))\n",
        "  print(f'The index of this highest score is {index}')\n",
        "  best_for_accuracy = c_s[index]\n",
        "  print(f'Therefore, the best C value is {best_for_accuracy}')\n",
        "  return best_for_accuracy"
      ],
      "metadata": {
        "id": "PTSyKHmWCgKH"
      },
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "best_for_ac1 = find_accuracy_c_value(accs1, c_values)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5TFzcFYACvS_",
        "outputId": "230a10bf-ad85-4898-8731-1bed922b483c"
      },
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The maximum accuracy achieved during 5-fold cross-val is 0.804001308538186\n",
            "The index of this highest score is 22\n",
            "Therefore, the best C value is 0.029470517025518096\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Using above C-value, train and test the selected (accurate) model\n",
        "\n",
        "def test_final_model_accuracy(best_for_accuracy, X_train, y_train, X_test, y_test, reweighed=False):\n",
        "  # only fit Classifier on the training data\n",
        "  clf = LogisticRegression(C=best_for_accuracy, solver='liblinear', random_state=1)  \n",
        "  if reweighed:\n",
        "    clf.fit(X_train,y_train, sample_weight=train.instance_weights)\n",
        "  else:\n",
        "    clf.fit(X_train,y_train)\n",
        "  # get predictions \n",
        "  pred = clf.predict(X_test)\n",
        "  # compute the accuracy\n",
        "  accuracy = clf.score(X_test, y_test)\n",
        "  print(f'The final accuracy for this model is {accuracy}')\n",
        "  return accuracy\n",
        "\n",
        "best_accuracy = test_final_model_accuracy(best_for_ac1, X_train, y_train, X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Cfi-QFbsDKHb",
        "outputId": "8f5e9c18-b4ec-4dd8-a1f6-764df5e758d3"
      },
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The final accuracy for this model is 0.8042039172865625\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#best_accuracy = test_final_model_accuracy(6.25E-06, X_train, y_train, X_test, y_test,reweighed=False)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "N4fjB5KaOnaw",
        "outputId": "d344837d-5671-4a75-9a58-104898dcc0e5"
      },
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The final accuracy for this model is 0.64\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Step 3. Fairness parameter sweep"
      ],
      "metadata": {
        "id": "P-ygJNn1jeCW"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def perform_cross_fold_fairness(folds, training_data, reg_strength, train_weights=None, reweighed=False): \n",
        "  \"\"\" \n",
        "    only pass in X_train and y_train into this function as we only want to \n",
        "  perform cross validation on the training \n",
        "  \"\"\"\n",
        "\n",
        "  eq_ops = []\n",
        "\n",
        "  # first convert into a subset object\n",
        "  iterable_data = training_data.copy()\n",
        "\n",
        "  # loop through fold dictionary and train and test\n",
        "  for fold_num, indices in folds.items():\n",
        "\n",
        "    train, test = iterable_data.subset(indices[0]), iterable_data.subset(indices[1])\n",
        "\n",
        "    X_train = train.features\n",
        "    y_train = train.labels.ravel()\n",
        "    X_test = test.features\n",
        "    y_test = test.labels.ravel() \n",
        "    #print(X_train.shape, y_train.shape, X_test.shape, y_test.shape )\n",
        "\n",
        "    # define classifier\n",
        "    clf = LogisticRegression(C=reg_strength, solver='liblinear', random_state=1)  \n",
        "\n",
        "    # get training reweights if relevant\n",
        "    if reweighed == False:\n",
        "      sample_weights = None\n",
        "    else:\n",
        "      sample_weights = train_weights[indices[0]]\n",
        "\n",
        "    # train classifier on training data\n",
        "    clf.fit(X_train, y_train, sample_weight=sample_weights)\n",
        "\n",
        "    # get predictions \n",
        "    pred = clf.predict(X_test)\n",
        "\n",
        "    # compute equality of opportunity (fairness) metric\n",
        "    test_pred = test.copy()\n",
        "    test_pred.labels = pred\n",
        "\n",
        "    metric = ClassificationMetric(test, test_pred, unprivileged_groups=unprivileged_groups, privileged_groups=privileged_groups)\n",
        "    eq_op = metric.equal_opportunity_difference()\n",
        "\n",
        "    # append to list\n",
        "    eq_ops.append(eq_op)\n",
        "  \n",
        "  # return the mean and standard deviation of the fairness metrics\n",
        "  return np.mean(eq_ops), np.std(eq_ops)"
      ],
      "metadata": {
        "id": "y4-Hop0rjgv6"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def parameter_sweep_fairness(c_s, reweighed=False):\n",
        "\n",
        "  eq_ops = []\n",
        "  stds = []\n",
        "\n",
        "  # loop through C values and perform 5-fold cross validation on training dataset\n",
        "  for c_value in c_s:\n",
        "    if reweighed:\n",
        "      eq_op, std = perform_cross_fold_fairness(folds, train, reg_strength=c_value, train_weights=train.instance_weights, reweighed=True)\n",
        "    else:\n",
        "      eq_op, std = perform_cross_fold_fairness(folds, train, reg_strength=c_value)\n",
        "\n",
        "    eq_ops.append(eq_op)\n",
        "    stds.append(std)\n",
        "\n",
        "  return eq_ops, stds  "
      ],
      "metadata": {
        "id": "VXjYhHXnjje1"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# perform hyperparameter sweep\n",
        "eq_ops1, stds_fair1 = parameter_sweep_fairness(c_values)"
      ],
      "metadata": {
        "id": "N_sEd6Pojlor"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plot_graph(c_values, eq_ops1, stds_fair1, 'fairness', 'Adult')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "MEbWVbu8jnOA",
        "outputId": "34542c93-c51b-4dd9-85b1-1b3698d3726c"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def find_fairness_c_value(equality_of_ops, c_s):\n",
        "  index1 = equality_of_ops.index(min(equality_of_ops, key=abs))\n",
        "  print(min(equality_of_ops, key=abs))\n",
        "  print(f'The index of this best score is {index1}')\n",
        "  best_cs = c_s[index1]\n",
        "  print(f'Therefore, the best C value is {best_cs}')\n",
        "  return best_cs\n",
        "\n",
        "# get the best model\n",
        "best_for_fair1 = find_fairness_c_value(eq_ops1, c_values)\n",
        "print(best_for_fair1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "whn5xPrtEYDD",
        "outputId": "8cdbe8e5-4a6b-4426-f58d-a52cdaec1089"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "-0.2324529173619491\n",
            "The index of this best score is 16\n",
            "Therefore, the best C value is 6.250551925273976e-06\n",
            "6.250551925273976e-06\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Using above C-value, train and test the selected (fair) model\n",
        "\n",
        "def test_final_model_fairness(best_fair, test, X_train, y_train, X_test, y_test, reweighed=False):\n",
        "  # only fit Classifier on the training data\n",
        "  clf = LogisticRegression(C=best_fair, solver='liblinear', random_state=1)  \n",
        "  if reweighed:\n",
        "    clf.fit(X_train,y_train, sample_weight=train.instance_weights)\n",
        "  else:\n",
        "    clf.fit(X_train,y_train)\n",
        "  # get predictions \n",
        "  pred = clf.predict(X_test)\n",
        "  # compute equality of opportunity (fairness) metric\n",
        "  test_pred = test.copy()\n",
        "  test_pred.labels = pred\n",
        "  metric = ClassificationMetric(test, test_pred, unprivileged_groups=unprivileged_groups, privileged_groups=privileged_groups)\n",
        "  eq_op = metric.equal_opportunity_difference()\n",
        "  print(f'The final equality of opportunity value for this model is {eq_op}')\n",
        "\n",
        "  return eq_op"
      ],
      "metadata": {
        "id": "nZjIJnfgEneE"
      },
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "final_fairness = test_final_model_fairness(best_for_fair1, test, X_train, y_train, X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "m7-qXRWIEpnt",
        "outputId": "b7ab2309-c722-479a-81c1-7df120be29a9"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The final equality of opportunity value for this model is -0.24153321836248665\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#final_fairness = test_final_model_fairness(0.029470517025518096, test, X_train, y_train, X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "KHFYkmhZPNpE",
        "outputId": "a5a3358d-eec8-4190-c7b4-d8e0633565cb"
      },
      "execution_count": 83,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The final equality of opportunity value for this model is -0.4632554945054945\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Step 4. Reweigh the data\n"
      ],
      "metadata": {
        "id": "HkPZn81Yjxp3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# transforming the dataset by reweighing.\n",
        "RW = Reweighing(unprivileged_groups=unprivileged_groups, privileged_groups=privileged_groups)\n",
        "\n",
        "# this gives us instance weights\n",
        "train = RW.fit_transform(train)\n",
        "\n",
        "# 4 groups: male 50k+, female 50k+, male < 50k, female < 50k\n",
        "print(\"subgroup weights\", np.unique(train.instance_weights)) "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Wy8oAgC0j78A",
        "outputId": "0e62afd9-31af-4778-8e8b-ff3256e80cbf"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "subgroup weights [0.78875735 0.85514075 1.09270553 2.1493453 ]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Step 5. Accuracy parameter sweep: Reweighed"
      ],
      "metadata": {
        "id": "rh6APOPSkEh0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Rewighed!\n",
        "accs2, stds2 = parameter_sweep_accuracy(c_values, reweighed=True)\n"
      ],
      "metadata": {
        "id": "6VRN0ZWXj9sA"
      },
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plot_graph(c_values, accs2, stds2, 'accuracy', 'Adult')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "1AB-xhDJj_cc",
        "outputId": "25ea045c-12b0-4f44-fca1-cee7b179e35f"
      },
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_for_ac2 = find_accuracy_c_value(accs2, c_values)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ni7cJ_x5FVmv",
        "outputId": "adec0970-d754-4b9e-c741-be0b35faab8d"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The maximum accuracy achieved during 5-fold cross-val is 0.7896985173023461\n",
            "The index of this highest score is 18\n",
            "Therefore, the best C value is 0.00010481131341546875\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_accuracy2 = test_final_model_accuracy(best_for_ac2, X_train, y_train, X_test, y_test,reweighed=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "klMOu5iRFSWC",
        "outputId": "fca5f6bb-e9f8-49fa-e65e-aa66178b22db"
      },
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The final accuracy for this model is 0.7905548351873336\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#best_accuracy2 = test_final_model_accuracy(2.559547922699533e-05, X_train, y_train, X_test, y_test,reweighed=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WEIUMXJ4Pvzw",
        "outputId": "9270f6d9-4c0f-4efe-ad3a-e24ce6d42b56"
      },
      "execution_count": 85,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The final accuracy for this model is 0.6366666666666667\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Step 6. Fairness parameter sweep: Reweighed"
      ],
      "metadata": {
        "id": "7YskGtCRkILJ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "eq_ops2, stds_fair2 = parameter_sweep_fairness(c_values, reweighed=True)"
      ],
      "metadata": {
        "id": "FxAJ1V_lkNOX"
      },
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plot_graph(c_values, eq_ops2, stds_fair2, 'fairness', 'Adult')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "qvBGyod_kQCY",
        "outputId": "aba975be-68f4-4092-b3ea-f9156fe7c118"
      },
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_for_fair2 = find_fairness_c_value(eq_ops2, c_values)"
      ],
      "metadata": {
        "id": "JSKBgP7qnd-9",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "75ae7c60-c842-41f9-e61c-f685553b1b42"
      },
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "-0.00035721253218468085\n",
            "The index of this best score is 17\n",
            "Therefore, the best C value is 2.559547922699533e-05\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "final_fairness2 = test_final_model_fairness(best_for_fair2, test, X_train, y_train, X_test, y_test,reweighed=True)"
      ],
      "metadata": {
        "id": "QyPxp01Bkene",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5050d81d-edba-48cb-eff9-6b1f1499b647"
      },
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The final equality of opportunity value for this model is 0.035579370945224564\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#final_fairness2 = test_final_model_fairness(0.00010481131341546875, test, X_train, y_train, X_test, y_test,reweighed=True)"
      ],
      "metadata": {
        "id": "400F6jglkeqU",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b7361ade-81bd-4061-c1f4-52db186b2da3"
      },
      "execution_count": 84,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The final equality of opportunity value for this model is -0.15521978021978022\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "M3ctclp2ke0M"
      },
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "R8MeVyZ6ke57"
      },
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "Js_WGMhuke-2"
      },
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "P-FcuZZEkfDI"
      },
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "ykBXAzJbkfGU"
      },
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#GERMAN\n",
        "\n",
        "##Step 1. Prepare the dataset"
      ],
      "metadata": {
        "id": "869-ZIWmkT6q"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# label the sensitive feature to be investigated\n",
        "privileged_groups = [{'age': 1}]\n",
        "unprivileged_groups = [{'age': 0}]\n",
        "\n",
        "# load in the dataset\n",
        "dataset_orig = load_preproc_data_german(['age'])\n",
        "\n",
        "# split training and test data\n",
        "train, test = dataset_orig.split([0.7], shuffle=True)\n",
        "\n",
        "# normalise dataset\n",
        "ss = StandardScaler()\n",
        "# try actually normalising the whole 'train' and 'test' dataset objects\n",
        "train.features = ss.fit_transform(train.features)\n",
        "X_train = train.features\n",
        "y_train = train.labels.ravel()\n",
        "test.features = ss.transform(test.features) \n",
        "X_test = test.features\n",
        "y_test = test.labels.ravel()"
      ],
      "metadata": {
        "id": "LlB8ci6QkbUT"
      },
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# create folds dict\n",
        "folds = get_folds(X_train)"
      ],
      "metadata": {
        "id": "3SrpzjNNmOlC"
      },
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Step 2. Accuracy parameter sweep"
      ],
      "metadata": {
        "id": "29ovdpaHmc23"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# perform hyperparameter sweep over reg strengths\n",
        "accs1_G, stds1_G = parameter_sweep_accuracy(c_values)"
      ],
      "metadata": {
        "id": "chIR-Pl2mag9"
      },
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plot_graph(c_values, accs1_G, stds1_G, 'accuracy', 'German')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "YmPB3XWOmqvg",
        "outputId": "5ebb61d8-9b0f-4621-e8b0-b549adb564a1"
      },
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_for_ac1_G = find_accuracy_c_value(accs1_G, c_values)\n",
        "best_accuracy1_G = test_final_model_accuracy(best_for_ac1_G, X_train, y_train, X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7Iy7O7FmGTRF",
        "outputId": "3e6a2a56-53a6-4a84-f101-5215281cd4be"
      },
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The maximum accuracy achieved during 5-fold cross-val is 0.6957142857142857\n",
            "The index of this highest score is 22\n",
            "Therefore, the best C value is 0.029470517025518096\n",
            "The final accuracy for this model is 0.7033333333333334\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#best_accuracy1_G = test_final_model_accuracy(0.029470517025518096, X_train, y_train, X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0SXi1c6yP8HA",
        "outputId": "ae88a342-1e86-46ab-e7ae-68fccd2c0088"
      },
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The final accuracy for this model is 0.7033333333333334\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Step 3. Fairness parameter sweep \n"
      ],
      "metadata": {
        "id": "Z0J5MeZ5mym5"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# perform hyperparameter sweep\n",
        "eq_ops1_G, stds_fair1_G = parameter_sweep_fairness(c_values)"
      ],
      "metadata": {
        "id": "jAuijXrzm1gr"
      },
      "execution_count": 39,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plot_graph(c_values, eq_ops1_G, stds_fair1_G, 'fairness', 'German')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "UhURc5X-m7do",
        "outputId": "576234e1-dbbd-42e4-99bc-eb14af529966"
      },
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_for_fair1_G = find_fairness_c_value(eq_ops1_G, c_values)\n",
        "final_fairness1_G = test_final_model_fairness(best_for_fair1_G, test, X_train, y_train, X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "erz97YCAGd0E",
        "outputId": "eb381318-b09f-4be0-e83d-b192a93b93c4"
      },
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "-0.33646940627430644\n",
            "The index of this best score is 22\n",
            "Therefore, the best C value is 0.029470517025518096\n",
            "The final equality of opportunity value for this model is -0.4632554945054945\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#final_fairness1_G = test_final_model_fairness(0.029470517025518096, test, X_train, y_train, X_test, y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "fqvWnzREQJKM",
        "outputId": "e41af266-4533-4deb-9961-b6493f633bd1"
      },
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The final equality of opportunity value for this model is -0.4632554945054945\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Step 4. Reweigh"
      ],
      "metadata": {
        "id": "nBFBQNVKnApI"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# transforming the dataset by reweighing.\n",
        "RW = Reweighing(unprivileged_groups=unprivileged_groups, privileged_groups=privileged_groups)\n",
        "\n",
        "# this gives us instance weights\n",
        "train = RW.fit_transform(train)\n",
        "\n",
        "# 4 groups: male 50k+, female 50k+, male < 50k, female < 50k\n",
        "print(\"subgroup weights\", np.unique(train.instance_weights)) "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Um9AFxVUnFMs",
        "outputId": "7872d4fc-3250-4c2e-aa7e-859eafb0a26a"
      },
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "subgroup weights [0.71684729 0.9597479  1.10527473 1.21054945]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Step 5. Accuracy parameter sweep: reweighed"
      ],
      "metadata": {
        "id": "xPfopYgWnIyb"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Rewighed!\n",
        "accs2_G, stds2_G = parameter_sweep_accuracy(c_values, reweighed=True)\n"
      ],
      "metadata": {
        "id": "9swnyY48nIPQ"
      },
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plot_graph(c_values, accs2_G, stds2_G, 'accuracy', 'German')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "tE8Kf1brnSXN",
        "outputId": "d3d4e8d9-d26a-46b0-d862-02b239b92417"
      },
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_for_ac2_G = find_accuracy_c_value(accs2_G, c_values)\n",
        "best_accuracy2_G = test_final_model_accuracy(best_for_ac2_G, X_train, y_train, X_test, y_test,reweighed=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZxgvcjpEGvC1",
        "outputId": "dee5da6f-c02f-424e-adf8-9b9ba7bf0342"
      },
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The maximum accuracy achieved during 5-fold cross-val is 0.6928571428571428\n",
            "The index of this highest score is 21\n",
            "Therefore, the best C value is 0.007196856730011528\n",
            "The final accuracy for this model is 0.68\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#best_accuracy2_G = test_final_model_accuracy(0.029470517025518096, X_train, y_train, X_test, y_test,reweighed=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8toCqNC6Q2wt",
        "outputId": "24175ebc-5e51-43d9-bd6c-daaa582290a0"
      },
      "execution_count": 89,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The final accuracy for this model is 0.68\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "##Step 6. Fairness parameter sweep: Reweighed"
      ],
      "metadata": {
        "id": "X8R1JmQInV51"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "eq_ops2_G, stds_fair2_G = parameter_sweep_fairness(c_values, reweighed=True)"
      ],
      "metadata": {
        "id": "aFOH3yxCnaQb"
      },
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plot_graph(c_values, eq_ops2_G, stds_fair2_G, 'fairness', 'German')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "qmB9Cr0kngKy",
        "outputId": "a6412c9f-3b94-43c4-a964-d3809bfe4673"
      },
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "best_for_fair2_G = find_fairness_c_value(eq_ops2_G, c_values)\n",
        "final_fairness2_G = test_final_model_fairness(best_for_fair2_G, test, X_train, y_train, X_test, y_test,reweighed=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GZhm_7_WG1Ic",
        "outputId": "8eff6132-5eb6-4210-9997-e35526ef5bba"
      },
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.050281034642801134\n",
            "The index of this best score is 22\n",
            "Therefore, the best C value is 0.029470517025518096\n",
            "The final equality of opportunity value for this model is 0.08241758241758246\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#final_fairness2_G = test_final_model_fairness(0.007196856730011528, test, X_train, y_train, X_test, y_test,reweighed=True)"
      ],
      "metadata": {
        "id": "e1p7XpQWnqZR",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5e80fa67-4712-49d0-a4ef-365683b1f049"
      },
      "execution_count": 88,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "The final equality of opportunity value for this model is 0.08241758241758246\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "kloSNssznqb5"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "VRJjgOTynqe6"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "k4sJjvrWnqh_"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Plot German and Adult together"
      ],
      "metadata": {
        "id": "qNQkPg90nrRE"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# plot adult\n",
        "#plt.scatter(np.log10(c_values), accs1, label='Adult' )\n",
        "plt.errorbar(np.log10(c_values), accs1, yerr = stds1, fmt='.', ecolor = 'magenta', color='black', label='Adult')\n",
        "\n",
        "#plot german\n",
        "#plt.scatter(np.log10(c_values), accs1, label='German')\n",
        "plt.errorbar(np.log10(c_values), accs1_G, yerr = stds1_G, fmt='.', ecolor = 'cyan', color='blue', label='German')\n",
        "\n",
        "plt.title('Effect of regularisation strength on accuracy: standard model')\n",
        "\n",
        "\n",
        "# C is the inverse of regularization strength\n",
        "plt.xlabel(\" $Log_{10}$(C)\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend()\n",
        "plt.show()  "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "18izZeeknvYf",
        "outputId": "16df99ab-41b5-4ad2-9e83-f5fd01ae18ba"
      },
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# plot adult\n",
        "#plt.scatter(np.log10(c_values), accs1, label='Adult' )\n",
        "plt.errorbar(np.log10(c_values), eq_ops1, yerr = stds_fair1, fmt='.', ecolor = 'magenta', color='black', label='Adult')\n",
        "\n",
        "#plot german\n",
        "#plt.scatter(np.log10(c_values), accs1, label='German')\n",
        "plt.errorbar(np.log10(c_values), eq_ops1_G, yerr = stds_fair1_G, fmt='.', ecolor = 'cyan', color='blue', label='German')\n",
        "\n",
        "plt.title('Effect of regularisation strength on fairness: standard model')\n",
        "\n",
        "\n",
        "# C is the inverse of regularization strength\n",
        "plt.xlabel(\" $Log_{10}$(C)\")\n",
        "plt.ylabel(\"Equality of Opportunity\")\n",
        "plt.ylim(-0.6,0)\n",
        "plt.legend()\n",
        "plt.show() "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "DkKgLVKmqQEk",
        "outputId": "d521f179-468a-4bb0-c3ae-afc266ae8f6f"
      },
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# plot adult\n",
        "#plt.scatter(np.log10(c_values), accs1, label='Adult' )\n",
        "plt.errorbar(np.log10(c_values), accs2, yerr = stds2, fmt='.', ecolor = 'magenta', color='black', label='Adult')\n",
        "\n",
        "#plot german\n",
        "#plt.scatter(np.log10(c_values), accs1, label='German')\n",
        "plt.errorbar(np.log10(c_values), accs2_G, yerr = stds2_G, fmt='.', ecolor = 'cyan', color='blue', label='German')\n",
        "\n",
        "plt.title('Effect of regularisation strength on accuracy: fairness-based model')\n",
        "\n",
        "\n",
        "# C is the inverse of regularization strength\n",
        "plt.xlabel(\" $Log_{10}$(C)\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend()\n",
        "plt.show() "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "SWOEOJ5CqCzn",
        "outputId": "9903303a-62fb-4d0d-fb28-4620095de372"
      },
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# plot adult\n",
        "#plt.scatter(np.log10(c_values), accs1, label='Adult' )\n",
        "plt.errorbar(np.log10(c_values), eq_ops2, yerr = stds_fair2, fmt='.', ecolor = 'magenta', color='black', label='Adult')\n",
        "\n",
        "#plot german\n",
        "#plt.scatter(np.log10(c_values), accs1, label='German')\n",
        "plt.errorbar(np.log10(c_values), eq_ops2_G, yerr = stds_fair2_G, fmt='.', ecolor = 'cyan', color='blue', label='German')\n",
        "\n",
        "plt.title('Effect of regularisation strength on fairness: fairness-based model')\n",
        "\n",
        "\n",
        "# C is the inverse of regularization strength\n",
        "plt.xlabel(\" $Log_{10}$(C)\")\n",
        "plt.ylabel(\"Equality of Opportunity\")\n",
        "plt.legend(loc='lower right')\n",
        "plt.show() "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "id": "sJIisE7HrFGk",
        "outputId": "e39fe54a-616f-4ace-ae32-9311a529bf7a"
      },
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "UEOr5YHzrxgf"
      },
      "execution_count": 52,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# plot accuracy against fairness for a given c value\n",
        "\n",
        "plt.plot(eq_ops1, accs1, 'x')\n",
        "plt.xlim(-0.5, 0.1)\n",
        "plt.ylabel('Accuracy')\n",
        "plt.xlabel('Equality of opportunity')\n",
        "plt.title('Standard model')\n",
        "plt.show()\n",
        "\n",
        "plt.plot(eq_ops2, accs2, 'x')\n",
        "plt.xlim(-0.5, 0.1)\n",
        "plt.ylabel('Accuracy')\n",
        "plt.xlabel('Equality of opportunity')\n",
        "plt.title('Fairness-based model')\n",
        "plt.show()\n",
        "\n",
        "plt.plot(eq_ops1_G, accs1_G, 'x')\n",
        "plt.xlim(-0.5, 0.1)\n",
        "plt.ylabel('Accuracy')\n",
        "plt.xlabel('Equality of opportunity')\n",
        "plt.show()\n",
        "\n",
        "plt.plot(eq_ops2_G,accs2_G, 'x')\n",
        "plt.xlim(-0.5, 0.1)\n",
        "plt.ylabel('Accuracy')\n",
        "plt.xlabel('Equality of opportunity')\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "Yn18vJ1exaR6",
        "outputId": "cc9d7036-3084-4349-ba5c-ab0366f89cc3"
      },
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "HXJ1lA0NxaWZ"
      },
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "5WjM9WDExaYi"
      },
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "APvNsrvYxaal"
      },
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "cfd5w3woxac8"
      },
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "DgM44bLWxaew"
      },
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "ZY5XEAgExaf6"
      },
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def accurate_fair_combined_NONGLOBAL(accuracies, eq_opportunities):\n",
        "\n",
        "  # normalise accuracy data\n",
        "  accs_norm = np.array(accuracies)\n",
        "  accs_norm = accs_norm - np.amin(accs_norm)\n",
        "  accs_norm = accs_norm / (np.amax(accs_norm)-np.amin(accs_norm))\n",
        "  #accs_norm = accs_norm / (np.amax(accs_norm)-0)\n",
        "  accs_norm = 1 - accs_norm\n",
        "  \n",
        "  # normalise fairness data\n",
        "  eq_ops_norm = np.array(eq_opportunities)\n",
        "  eq_ops_norm = np.abs(eq_ops_norm)\n",
        "  eq_ops_norm = eq_ops_norm - np.min(eq_ops_norm)\n",
        "  eq_ops_norm = eq_ops_norm / (np.max(eq_ops_norm)-np.min(eq_ops_norm))\n",
        "\n",
        "  fair_and_ac = accs_norm + eq_ops_norm\n",
        "\n",
        "  # do I normalise the output too?!! How do I make it comparable between datasets / standard vs fair?\n",
        "  #fair_and_ac = fair_and_ac - np.min(fair_and_ac)\n",
        "  #fair_and_ac = fair_and_ac / (np.max(fair_and_ac)- np.min(fair_and_ac))\n",
        "\n",
        "  return fair_and_ac, np.std(fair_and_ac)\n"
      ],
      "metadata": {
        "id": "6bIe2KPhrxjg"
      },
      "execution_count": 54,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def accurate_fair_combined(accuracies, eq_opportunities, accuracies2, eq_opportunities2):\n",
        "\n",
        "  # find global max range\n",
        "  accs_norm = np.array(accuracies)\n",
        "  accs_norm2 = np.array(accuracies2)\n",
        "\n",
        "  range1 = np.amax(accs_norm)-np.amin(accs_norm)\n",
        "  range2 = np.amax(accs_norm2)-np.amin(accs_norm2)\n",
        "\n",
        "  if range1 > range2:\n",
        "    range = range1\n",
        "    min1 = np.amin(accs_norm)\n",
        "  else:\n",
        "    range = range2\n",
        "    min1 = np.amin(accs_norm2)\n",
        "\n",
        "  eq_ops_norm = np.array(eq_opportunities)\n",
        "  eq_ops_norm2 = np.array(eq_opportunities2)\n",
        "\n",
        "  erange1 = np.amax(eq_ops_norm)-np.amin(eq_ops_norm)\n",
        "  erange2 = np.amax(eq_ops_norm2)-np.amin(eq_ops_norm2)\n",
        "\n",
        "  if erange1 > erange2:\n",
        "    erange = erange1\n",
        "    min2 = np.amin(eq_ops_norm)\n",
        "  else:\n",
        "    erange = erange2\n",
        "    min2 = np.amin(eq_ops_norm2)\n",
        "\n",
        "  # normalise accuracy data\n",
        "  accs_norm = accs_norm - min1\n",
        "  accs_norm = accs_norm / range\n",
        "  accs_norm = 1 - accs_norm\n",
        "  \n",
        "  # normalise fairness data\n",
        "  eq_ops_norm = np.abs(eq_ops_norm)\n",
        "  eq_ops_norm = eq_ops_norm - min2\n",
        "  eq_ops_norm = eq_ops_norm / erange\n",
        "\n",
        "  fair_and_ac = accs_norm + eq_ops_norm\n",
        "\n",
        "  # do I normalise the output too?!! How do I make it comparable between datasets / standard vs fair?\n",
        "  #fair_and_ac = fair_and_ac - np.min(fair_and_ac)\n",
        "  #fair_and_ac = fair_and_ac / (np.max(fair_and_ac)- np.min(fair_and_ac))\n",
        "\n",
        "  return fair_and_ac, np.std(fair_and_ac)"
      ],
      "metadata": {
        "id": "neI1X5iQWlua"
      },
      "execution_count": 55,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "fair_and_accurate1, stds_both1 = accurate_fair_combined_NONGLOBAL(accs1, eq_ops1)\n",
        "plot_graph(c_values, fair_and_accurate1, stds_both1, 'both', 'adult')\n",
        "\n",
        "np.where(fair_and_accurate1 == np.min(fair_and_accurate1))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 316
        },
        "id": "FtAJ3vEZrxmx",
        "outputId": "23739b90-c126-4c3d-e87c-312bc392f1b7"
      },
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEaCAYAAAAL7cBuAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de5hcVZ3u8e/bt3QHknBJBEOuICpBrrZgRoUeUQmckahzAy+Do0d0jjgeR2dABxFRRs8cZ8bjEUUYEUVNREadHAZBCbaIRElHCBIUDIHcCBDoXIB00rff+WPvSnY61d3V3bW7qrvfz/PU01VrX9baVdX7V2utvddSRGBmZtZXTaULYGZm1ckBwszMinKAMDOzohwgzMysKAcIMzMrygHCzMyKcoDIiaTPSnpG0pPp67dK2ijpeUmnVLBcVVGOtCw3SPrsCLb/saQLy1mmPPc7lkh6t6S7K5DvPEkhqa5M+7tC0rfLsa8S8mqRtKnc61aSA8QwSXpcUkd6oi08vpwumwN8FFgQEUemm3wBuDgiDo6I+0aQb0h6yQiKXpZyVIOIOCcivjmSfRQ7gZRjv0MsQ0VPFuU+KVdKpd/HkUjPJ2+odDn6GtNfiCrw5oi4o0j6HODZiHg6kzYXWDM6xRpQyeWQVBcR3TmXZ8gkCVBE9Fa6LKOlWj8LG99cgyiz9FfAT4GZaa1iiaTngVpgtaRH0/VmSvoPSVslPSbpbzP7qJX0CUmPSnpO0ipJsyXdla6yOt33XxbJv0bSZZLWS3pa0rckTZM0qVg5imwfkj4o6Q/AH9K0P5F0v6Ttku6RdGJm/VMl3ZeW8/uSvldoNirWTNFfDUjSoZJuSd+PbenzWZnlrZKukvRLYBdwdJr239PlL5H0c0k70qa972W2/T9ps9rO9L18XZq+CPgE8Jfp+7k6k1dhv0Xfz3RZ4Zf3hZI2pPn+Y7H3NV3/XEkPpe/VZkkfk3QQ8GP2fV+eT78bV0i6WdK3Je0E3p1+jl+XtCXd/rOSarPvtaQvpO/fY5LOyeQ9X9Jdad53SLpa+2pOhe/V9jT/hZntiu6vyLEdl75v2yWtkXReZtkNaX7/leb/a0nH9Lev1HskPZEe68cy+5ok6YvpsifS55P6ex/TzRrSz+25tGzNg+SdPa6i3510WVN6bNskPQS8qs+2+33X1U+TqqQbSX5U/r+03P9QavlyFxF+DOMBPA68oZ9lLcCmPmkBvCR9XgOsAi4HGoCjgXXA2enyvwd+C7wMEHAScHjf/fST93uAtek+DwZ+ANxYrBz9bB8kAe4woAk4BXgaOJ0kuFyYHvuktOzrgQ8D9cDbgE7gs+m+3g3cPcD7cENm3cOBPwUmA1OA7wM/ymzXCmwAjiep+danaf89Xb4E+Mf0vW0EXpvZ9p3p/utImv6eBBrTZVcA3+5Txux++30/gXnp8VyXvlcnAXuA4/p5b7cAr0ufHwqcOsD35QqgC3hLekxNwA+BrwEHAS8C7gXen3mvu4D3pZ/T3wBPkNS0AFaQNC82AK8FdhaOO3McdZn8B9xfn7LWp+/RJ9L9vx54DnhZ5nN+Fjgt/Qy+Ayzt5z0qlGVJepwnAFtJ/9eAK4Ffpcc/A7gH+Mwg7+Nu4Nz0OD4H/Cqz/CvAVwb4fxjou/N54Bck/yuzgQez+dPnf439v+8tfdZ9nH7OJxU9z1W6AGP1kX6gzwPbM4/3Ffvw+35ZSE62G/os/zjwjfT5w8DifvId7AS/HPgfmdcvS//R60rcPoDXZ15/tfAPmEl7GDgTOAPYnD1pAHczjABRpBwnA9syr1uBK/us08q+E/m3gGuBWSV8dtuAk9LnVzBwgOj3/WTfyWxWZvm9wPn95LsBeD8wtU96se/LFcBdmddHkASfpkzaBcDPMu/12syyyWnZjiT5ddoNTM4s/zaDB4ii+ytyXK8jOXHWZNKWAFdkPud/zyw7F/h9P+9RoSwvz6T9M/D19PmjwLmZZWcDjw/yPt6Reb0A6BjsO1Lid2cdsCiz7CLGWYBwE9PIvCUiDsk8ritxu7kkVeHthQfJr68j0uWzSf4RhmMmya/6gvUkJ7Mjiq9e1MY+Zf1on7LOTvOZCWyO9BteZNuSSZos6WtpU85OkmaPQwpNKCXs+x9Ialv3ps0I78ns+2OSfpc2P20HpgHTSyxaKe/nk5nnu0hqGsX8KcnJcb2S5rCF/axX0PdzqAe2ZD6Hr5H8kj6gHBGxK316cHoM7Zm0vvvuT3/762smsDH27xNaDxxVbF8M/B4VK9/6NI9CXn0/j5kMrG/ejSqxQ36Q787MIuUcVxwgKmMj8Fif4DIlIs7NLB+sjbY/T5CcTAoKvx6fGsI++p7wr+pT1skRsYSkyeQoScqsPzvz/AWSX54ASDqS/n2U5Nf56RExlaR2AslJv1i59i9wxJMR8b6ImEnyK/0rSvolXkcSPP4CODQiDgF2ZPbb7z5T5Xg/C2VcGRGLSU7qPwJuGqQMfT+HPcD0zOcwNSKOLyHrLcBhkiZn0rKf02DvwWCeAGZLyp5P5pDULocrW745aR6FvPp+HoVlIz2O/ZTw3dlSpJxZu8h8/0lqc/0pa9nLxQGiMu4FnpN0SdrRVSvpFZIKnVz/DnxG0rFKnCjp8HTZUyTt4f1ZAnwk7ZQ8GPgn4Hsx/CtgrgM+IOn0tCwHSfpvkqaQtGv3ABdLqpO0mKSduWA1cLykkyU1klT3+zMF6CDpKD0M+NRQCinpz7WvU3sbyT9cb7rfbpJ27DpJlwNTM5s+Bczrc3LLKsv7KalB0jskTYuILpI+gMIv7qeAw5V2fhcTEVuAnwD/Immqks7zYySdOVjeEbEeaAOuSMuxEHhzZpWtaVkG+l4N5NckJ8N/kFQvqSXd/9Jh7g/gk2mt8njgr4HCRQdLgMskzZA0naQfr9DZPuj7OESDfXduAj6u5AKLWcCH+mx/P/D29P97EUmzbH8G+7+uCAeIkSlcdVB4/LCUjSKiB/gTknb2x4BnSIJC4Yv9ryRfvp+QnEi+TtJJCclJ9ptpM8NfFNn99cCNJE00j5F00PX94pYsItpIOiq/THLiXUvSPk1EdJJ0TL+XpA/mncAtJL90iYhHSDoV7yC5ImqgG6++SHKMz5B0Qt42xKK+Cvi1kiu1lgEfjoh1wO3pvh4haQLYzf7NAt9P/z4r6TdF9lvO9/NdwONpE9oHgHcARMTvSU5869LPtb8mk78i6QR+iOSzuBl4cYl5vwNYSNJZ/FmSE27hc9oFXAX8Ms3/1UM5qPR78GbgHJLP7yvAX6XHNVw/J/muLQe+EBE/SdM/SxLsHiC5kOM3adpQ3se9JF0j6Zp+Fg/23fl0mv4Yyf/qjX22/zDJ+7Kd5P3/0QBF+RxJ4NuuzFVblVa4wsGsLCT9GrgmIr5R6bJY/5RcBvz7iBhSTc0mFtcgbEQknSnpyLSJ6ULgRIb+699yJulVaZNUTdrcsZiBf9Ga+U5qG7GXkTSHHURy2d+fpe3lVl2OJLmH43BgE/A3McaHWrH8uYnJzMyKchOTmZkVNW6amKZPnx7z5s2rdDHMzMaUVatWPRMRM4otGzcBYt68ebS1tVW6GGZmY4qkfu8AdxOTmZkV5QBhZmZFOUCYmVlRDhBmZlaUA4SZmRXlAGFmZkU5QJiZWVEOEGZmVpQDhE04LenDzAbmAEH/J4xypY9GHuMlbzOrHg4QZmZWlAOEmZkV5QBhZmZFOUCYmVlRDhBmqRbG9wUAzrt68hiNvMshtwAh6XpJT0t6sJ/lkvQlSWslPSDp1MyyHkn3p49leZXRzMz6l2cN4gZg0QDLzwGOTR8XAV/NLOuIiJPTx3n5FdHMzPqTW4CIiLuA9gFWWQx8KxK/Ag6R9OK8ymNmZkNTyT6Io4CNmdeb0jSARkltkn4l6S397UDSRel6bVu3bs2zrGZmE061dlLPjYhm4O3AFyUdU2yliLg2IpojonnGjKJzbpuZ2TBVMkBsBmZnXs9K04iIwt91QCtwymgXzsxsoqtkgFgG/FV6NdOrgR0RsUXSoZImAUiaDrwGeKiC5TQzm5Dq8tqxpCUkl+dOl7QJ+BRQDxAR1wC3AucCa4FdwF+nmx4HfE1SL0kA+3xEOECYmY2y3AJERFwwyPIAPlgk/R7ghLzKZWZmpanWTmozM6swBwgzMyvKAcLMzIpygDAzs6IcIMzMrCgHCKsqLXiuarNq4QBhZmZFOUCYmVlRDhBmZlaUA4SZmRXlAGETzo4VK9jwuc+xYsWKShfFrKo5QND/CaNc6aORx3jJO28rVqzggde/nsc++UnOOussBwmzAeQ2WN9Y0fqLu1n9xjcSXZ201DdwyZe+w0tOOJW1v/0Nq//2HSNOB8q2r3GXd3cXZzU0sHz5chYuXDg6n3drK717OiF62dPZSWtr66jlbTbWTPgAsXz5z4jO5ITR2dnJ//3OfzJt4SR2rPjPsqQDZdvXeM17NE/SLS0tqLae6OkiVEvj7FeMSr5mY1JEjIvHK1/5yhiOu+/+ZdQ0NQW1tdHY1BQ/vqM1nnlud/z4jtaypJdzX+MpbzU0BqqJpqamuOeee/Z+Hmemj7x09/TGke/833HEuz4cZ3z0mph/6S1xc9vGAfMeqExD3aZc6c57bOcxGnmXCmiLfs6ruZ2wgeuBp4EH+1ku4EskEwY9AJyaWXYh8If0cWEp+Q03QEREnHzPPTH/n/5pvxNVOdNHI4+xlvdxN90ah5zxV/EvNy7bL/3MyDdAbHthT8y95JY4+Rfr4vndXfH261bE3EtuiW/d89iEPZE479HPYzTyLlWlAsQZwKkDBIhzgR+ngeLVwK/T9MOAdenfQ9Pnhw6W30gCxJkxdj708ZL3Gd09Mf/Tt8ffLvlNyfsqh8efeT7mXnJLnLpyQ0REdHR2x3tvWBlzL7klXvLlm8ZF8HXe1Z/HaORdqooEiCRf5g0QIL4GXJB5/TDwYuAC4Gv9rdffwwFi7OV9/PdXxysuvy12d3WXtK9yeGDj9ph7yS3xqge37E3r7O6JP//0N0J1DUFNTdQ3NMb//NL34uu/WBef+fcfhhobg5ramNTYGF9eckv87PdPxc9+/1R8ecktRZflne68x3YeeeXdt7m2VAMFiEp2Uh8FbMy83pSm9Zd+AEkXARcBzJkzJ59SWm4mv+JInm7byD1rn+WPX/6iUclzR0cXADVN9XvT6mtrOKnuCb7f2w29vXR1dXLDzf/FtM0HsWPFD4nCVU97gsu+ehPT0v70HStuKros73TnPbbzyCvvXC746C9ylOPBwDWIW4DXZl4vB5qBjwGXZdI/CXxssLxcgxh7eZ/R1R3HX35b/MP3V5e0r3K4ZfUTMfeSW2Lhlh37pd9zzz17O9Sbmprirl/cHdte2BM/vO3OqGlM0ic1NsU3fnBbrFrfHqvWt8c3fnBb0WV5pzvvsZ1HXnnnUYNwE1NU/kQ5kfP+0Hd/E6dc+ZPo6u4ZdF/l8N1fr4+5l9wSr9m264BlE7Wt2nmPz+Mr1UABQsnyfEiaB9wSEQdcbC7pvwEXk3RWnw58KSJOk3QYsIqkgxvgN8ArI6J9oLyam5ujra1tWOVsSf+25pQ+GnmM1bwv+e0W/uY7v+G77zudPzpm+oD7Kodrfv4on//x75nz6bO5a9L+Laz95T1QmYa6TbnSnffYzmM08i6VpFUR0VxsWW59EJKWkJR9uqRNwKeAeoCIuAa4lSQ4rAV2AX+dLmuX9BlgZbqrKwcLDjZ2nfmyGUyqq+H2B5/kj46Znnt+Ozq6oEaooTb3vMzGutwCRERcMMjyAD7Yz7LrSe6jsHFuckMdZ750BreveYpPvfl4qFGu+e3o6KKmqR4p33zMxgMP1mcVd84JR/Lkzt3cv2l77nntTAOEmQ3OAcIq7vUvP4K6GnH7g0/mntcOBwizkjlAWMVNa6rnj14yndvWPEmeF02AaxBmQ+EAYVVh0fFHsv7ZXXQ9+Vyu+ezo6KLWAcKsJA4QVhXedPwRSPBCzs1MO3d3U9M04Ue5NyuJA4RVhekHT+JV8w5jV44BIiKSPohG1yDMSuEAYVVj0fFH0vXUc3RtfT6X/b/Q2UNPb7gPwqxEDhBWNRa94kj2bP4dj33iilzmii42UJ+Z9c8BwqrG+t/dz1Pfu4wtX/83zjrrrLIHiZ0OEGZD4gBhVaO1tZXo6dpv6OJycg3CbGgcIKxqtLS0oLoGUA0NDQ20tLSUdf8OEGZD4wBhVWPhwoXM/9z1HHLmu7jjjjvKO/EJDhBmQ+ULwq2qTDnlVfQ8PZVTX3V62fftPgizoXENwqpKTWPym2Xn7q6y73tnRxcS1Ezy7yKzUjhAWFUp/Lp/bnd32fe9o6OLKZPqUM5DipuNFw4QVlX21iA6yl+D2NHRxbTJbl4yK1WuAULSIkkPS1or6dIiy+dKWi7pAUmtkmZllvVIuj99LMuznFY9CsNg5FGD2Lm7m2lN9bSS35SmZuNJnlOO1gJXA28ENgErJS2LiIcyq30B+FZEfFPS64HPAe9Kl3VExMl5lS+rNef00chjvOSdZ4DY0dHFVI/DZFayPGsQpwFrI2JdRHQCS4HFfdZZANyZPv9ZkeU2weTZSb2jo4tpvoLJrGR5BoijgI2Z15vStKzVwNvS528Fpkg6PH3dKKlN0q8kvaVYBpIuStdp27p1aznLbhWyr5PaAcKs0irdSf0x4ExJ9wFnApuBnnTZ3IhoBt4OfFHSMX03johrI6I5IppnzJgxaoW2/KihFgQ7O3Log3CAMBuSPC8I3wzMzryelabtFRFPkNYgJB0M/GlEbE+XbU7/rpPUCpwCPJpjea0KSKJmUl3ZaxC7u3rY093LVAcIs5IprzmAJdUBjwBnkQSGlcDbI2JNZp3pQHtE9Eq6CuiJiMslHQrsiog96TorgMV9Orj309zcHG1tbbkci42eFmDT/7qT8+Ydxr/+ZfmuUXh6525O+6flfOYtr+Bdr55btv2ajXWSVqWtNQfIrYkpIrqBi4Hbgd8BN0XEGklXSjovXa0FeFjSI8ARwFVp+nFAm6TVJJ3Xnx8oONj4osb6sndSF8ZhchOTWelyHXMgIm4Fbu2Tdnnm+c3AzUW2uwc4Ic+yWfWqaaxjZ5kvcy0EHAcIs9JVupPa7AA1jfVlvw+iUIOY2uhxmMxK5QBhVaemsa7sQ224icls6BwgrOrUNNWX/SqmHbscIMyGygHCqk5NYx3P7emmt7d8V9gV+jR8matZ6RwgrOrUNNYRAS90lq8fYkdHF5Mbaqmv9VferFT+b7Gqk8eAfR5mw2zoHCCs6hQCRDnvhXCAMBs6BwirOjVNyaWo5axB7Ozocv+D2RA5QFjV2dfEVN4ahOeCMBsaBwirOvumHS1vDcJNTGZD4wBhVSevGoQDhNnQOEBY1dHeWeXKU4Po7unlhc4eBwizIfLANFZVWgHqa3lpbU3ZrmLad5Ocv+5mQ+EahFWlqU11ZbuKyeMwmQ2PA4RVpSmN9WUbsM8Bwmx4cg0QkhZJeljSWkmXFlk+V9JySQ9IapU0K7PsQkl/SB8X5llOqz5TG8tXg9jpAGE2LLkFCEm1wNXAOcAC4AJJC/qs9gXgWxFxInAl8Ll028OATwGnA6cBn0qnIbUJYkpj+UZ03TsXhAOE2ZDkWYM4DVgbEesiohNYCizus84C4M70+c8yy88GfhoR7RGxDfgpsCjHslqVmVLGWeXcxGQ2PHkGiKOAjZnXm9K0rNXA29LnbwWmSDq8xG2RdJGkNkltW7duLVvBrfKm5lCDcIAwG5pKd1J/DDhT0n3AmcBmoKfUjSPi2ohojojmGTNm5FVGq4ApjXVlu5N65+4uGupqaKyvLcv+zCaKPAPEZmB25vWsNG2viHgiIt4WEacA/5imbS9lWxvfpjTW09HVQ1dP74j35WE2zIYnzwCxEjhW0nxJDcD5wLLsCpKmSyqU4ePA9enz24E3STo07Zx+U5pmE0Thprbny9APkQzU55vkzIYqtwAREd3AxSQn9t8BN0XEGklXSjovXa0FeFjSI8ARwFXptu3AZ0iCzErgyjTNJogpZZwTwuMwmQ1Prj+rIuJW4NY+aZdnnt8M3NzPttezr0ZhE0zhF3857oXY2dHN9IMbRrwfs4mm0p3UZkXtrUGU4W5q1yDMhscBwqrSlDKO6LrDs8mZDYsDhFWlwi/+kd4L0dsb7NztGoTZcDhAWFUqVw3i+c5uInyTnNlwOEBYVTp4UqGTemQ1iB27PA6T2XA5QFhVqqut4aCG2hFfxbR3oL5GBwizofLdQ1a1RjonRAvQ4XGYzIbNNQirWuWYVa53twOE2XANGCAknT3Asj8vf3HM9pnSWD/iO6l7CzWIyQ4QZkM1WA3iVkk/k3TAUNskYyeZ5WZKGWaV693bB+HWVLOhGixAPAB8F/iVpD/rs0z5FMksUY45IXo7uqit0d6rosysdIMFiIiI64CzgEskfUPS5MKyfItmE105ZpXr7ehmamMdkn/PmA1VSZ3UEfEIsBB4CrhP0um5lsqMffNSRwz/t0ivx2EyG7bB6t17f3alw3dfKuk2YAngKdwsV1Ob6ujqCfZ09w57Nrhej8NkNmyDBYhP902IiFZJrwTen0+RzBLZEV2HGyB6XIMwG7YBm5gi4kf9pG+LiM8PtnNJiyQ9LGmtpEuLLJ+TXiV1n6QHJJ2bps+T1CHp/vRxTakHZOPH1DKMx9S72zUIs+HK7dIOSbXA1cAbgU3ASknLIuKhzGqXkcw091VJC0gmF5qXLns0Ik7Oq3xW/aaWYVY590GYDV+ed1KfBqyNiHUR0QksBRb3WSeAqenzacATOZbHxpgpI5xVLiKSPgiPw2Q2LHkGiKOAjZnXm9K0rCuAd0raRFJ7+FBm2fy06ennkl5XLANJF0lqk9S2devWMhbdqsHUEc4JEV090BOuQZgNU6XHYroAuCEiZgHnAjdKqgG2AHMi4hTg74DvSprad+OIuDYimiOiecYMX1Q13uydE6JjeDWI3nQ7Bwiz4ckzQGwGZmdez0rTst4L3AQQESuARmB6ROyJiGfT9FXAo8BLcyyrVaHCVUzDrUH0eiRXsxHJM0CsBI6VNF9SA3A+sKzPOhtI7tJG0nEkAWKrpBlpJzeSjgaOBdblWFarQgc11FKj4fdB7B2HqcnDbJgNR27/ORHRLeli4HagFrg+ItZIuhJoi4hlwEeB6yR9hKTD+t0REZLOAK6U1AX0Ah+IiPa8ymrVSdKIRnR1DcJsZHL9aRURt5J0PmfTLs88fwh4TZHt/gP4jzzLZmPDSOaEcIAwG5lKd1KbDWjKpOHPKucAYTYyDhBW1UYyJ0RPGiCm+D4Is2FxgLCqNrVpZH0QmlRHbY2H+jYbDgcIq2ojqUH07u6ixs1LZsPmAGFVbeoIr2KqdYAwGzYHCKtqUxvreH5PN729Q580qLejixrfA2E2bA4QVtWmNNYTAc93Dr2Zqbej201MZiPgAGFVrXAX9HD6IZIahAOE2XA5QFhVy84qN1S7/vAAz/y/G1ixYkW5i2U2IThAWFUb7pwQd/3ilzz17Ut56ttf4qyzznKQMBsGBwiralOHOaLrj3+6nOjpht5eOjs7aW1tzaF0ZuObA4RVtUIN4pLdXbQMYbujTzwN1dZBbS0NDQ20tAxlazODnAfrMxupQh9E7xCbmKbNO54jzr+KhnkvsPTcN7Fw4cI8imc2rjlAWFUr1CCGGiA2tu9i0uwFzL1iEQtrXVE2Gw7/51hVa6yvpaGuZu/IrKXa0L6LukMakYOD2bDl+t8jaZGkhyWtlXRpkeVzJP1M0n2SHpB0bmbZx9PtHpZ0dp7ltOo2tbF+yDWIDe27qDtsck4lMpsYcgsQ6ZShVwPnAAuACyQt6LPaZcBNEXEKyZSkX0m3XZC+Ph5YBHylMAWpTTxTG+voHeJVTBsdIMxGLM8axGnA2ohYFxGdwFJgcZ91ApiaPp8GPJE+XwwsjYg9EfEYsDbdn01AUxrrhlSDeGFPN8883+kAYTZCeQaIo4CNmdeb0rSsK4B3StpEMjXph4awrU0QU5vqiSHUIDZu2wVAvQOE2YhUugfvAuCGiJgFnAvcKKnkMkm6SFKbpLatW7fmVkirrCmNdUPqpF7/bBIgXIMwG5k8A8RmYHbm9aw0Leu9wE0AEbECaASml7gtEXFtRDRHRPOMGTPKWHSrJlMmDa2TemO7A4RZOeQZIFYCx0qaL6mBpNN5WZ91NgBnAUg6jiRAbE3XO1/SJEnzgWOBe3Msq1WxqU1D64PY0L6LKY11HsnVbIRyu1EuIrolXQzcDtQC10fEGklXAm0RsQz4KHCdpI+QdFi/OyICWCPpJuAhoBv4YET05FVWq25TGuuJrh6ipxdKuK9hQ/su5hw2mRfkuajNRiLXO6kj4laSzuds2uWZ5w8Br+ln26uAq/Isn40NU7N3Ux/UMOj6G9p38bIjpvC7vAtmNs5VupPabFB7x2MqoaO6tzfY1N7BHPc/mI2YA4RVvaGMx/TUc7vp7OllzuEOEGYj5QBhVW9qU2FE18FrEBvSS1xdgzAbOQcIq3r7ahAlBIh2BwizcnGAsKo3dQhzQmxs30WNYOYhTXkXy2zcc4CwqjdQgGhJHwUb2ncx85Am6j3Mt9mI+b/Iqt7BhSamEq5iKtwDYWYj5wBhVa+2RmhSHVFCE9MGX+JqVjYOEDYm1JQwJ8Suzm6eeX4Psx0gzMrCAcLGhFICxMb2DsBXMJmViwOEjQk1JUw76ktczcrLAcLGhFJqEOuffQFwgDArFwcIGxNKqUFsbN/FlEl1HDLZw3yblYMDhI0Ju9evof0nN7JixYp+19nQvovZh01GHubbrCxyHe7brBxWrFjBY1e8j+js5KxfLmX58uUsXLjwgPU2tO/i2BdN2fu6dRTLaDYeuQZhVa+1tZXo6oLopbOzk9bW1gPW6e0NNm7r8CiuZmWUa4CQtEjSw5LWSrq0yPJ/k3R/+nhE0vbMsp7Msr5TldoE0tLSQs2kBlANNbX1tLS0HLDO08/tobO71/dAmJVRbk1MkmqBq4E3ApuAlZKWpbPIARARH8ms/yHglMwuOiLi5LzKZ2PHwoULOXH5cjb/y7eZMf3lnOuEDakAAAvISURBVH76qw9Yx5e4mpVfnjWI04C1EbEuIjqBpcDiAda/AFiSY3lsDJu2cCFzLvsELxxyNPc+3n7AcgcIs/LLM0AcBWzMvN6Uph1A0lxgPnBnJrlRUpukX0l6Sz/bXZSu07Z169Zylduq1OQFR9BUX8uy1U8csGxD+y4kOMrDfJuVTbV0Up8P3BwRPZm0uRHRDLwd+KKkY/puFBHXRkRzRDTPmDFjtMpqFVLTUMcbFhzBrb/dQmd3737LNrbvYua0JhrqquUrbTb25fnftBmYnXk9K00r5nz6NC9FxOb07zqSKxZPOXAzm2gWnzST7bu6uHvt/jXG5B4I1x7MyinPALESOFbSfEkNJEHggKuRJL0cOBRYkUk7VNKk9Pl04DXAQ323tYnnjJfOYFpTPcvu37+ZyfNAmJVfblcxRUS3pIuB24Fa4PqIWCPpSqAtIgrB4nxgaUREZvPjgK9J6iUJYp/PXv1kE1dDXQ3nnnAk/3n/E3R09kBDLb2dPWx9bo8DhFmZ5XondUTcCtzaJ+3yPq+vKLLdPcAJeZbNxq43nzSTJfdu5I7fPQUnzaR7W3oF0+EHVbhkZuOLe/RszDl9/uG8aMqkvVczdT/rS1zN8uAAYWNObY1480kz+fnDW+nZ1UW374Ewy4UDhI1J5500k86eXnat2UJX+y4OnlTHoR7m26ysPJqrjUknzprGvMMn8+TqJ1BdLfM8zLdZ2bkGYWOSJM47aSa7H32WPZt2MMf3QJiVnQOEjVnnnTwTAnqf9yWuZnlwgLAx6yUvmkLP7g3sWHETezb/vtLFMRt33AdhY9aKFSvYfM3fEZ2d/PO9N3HOCXcWnWnOzIbHNQgbs1pbW4nuZKa57q6uojPNmdnwOUDYmNXS0kJNQwPU1tLQ0FB0pjkzGz43MdmYVZhpbkdrK99paXHzklmZOUDYmDZt4UKmLVyIQ4NZ+bmJyczMinKAMDOzohwgzMysqFwDhKRFkh6WtFbSpUWW/5uk+9PHI5K2Z5ZdKOkP6ePCPMtpZmYHyq2TWlItcDXwRmATsFLSsuzMcBHxkcz6HyKdd1rSYcCngGYggFXpttvyKq+Zme0vzxrEacDaiFgXEZ3AUmDxAOtfACxJn58N/DQi2tOg8FNgUY5lNTOzPvK8zPUoYGPm9Sbg9GIrSpoLzAfuHGDbo4psdxFwEcCcOXNGXmKrWq2VLoDZBFQtndTnAzdHRM9QNoqIayOiOSKaZ8yYkVPRzMwmpjwDxGZgdub1rDStmPPZ17w01G3NzCwHeQaIlcCxkuZLaiAJAsv6riTp5cChwIpM8u3AmyQdKulQ4E1pmpmZjZLc+iAiolvSxSQn9lrg+ohYI+lKoC0iCsHifGBpRERm23ZJnyEJMgBXRkR7XmU1M7MDKXNeHtOam5ujra2t0sWwUdaS/m2tYBnMxjJJqyKiudiyaumkNjOzKuMAYWZmRTlAmJlZUQ4QZmZWlAOEmZkV5QBhZmZFOUCYmVlRDhBmZlaUA4SZmRXlAGFmZkU5QJiZWVEOEGZmVlSeM8qZ5a610gUwG8dcgzAzs6IcIMzMrKhcA4SkRZIelrRW0qX9rPMXkh6StEbSdzPpPZLuTx8HzERnZmb5yq0PQlItcDXwRmATsFLSsoh4KLPOscDHgddExDZJL8rsoiMiTs6rfGZmNrA8axCnAWsjYl1EdAJLgcV91nkfcHVEbAOIiKdzLI+ZmQ1BngHiKGBj5vWmNC3rpcBLJf1S0q8kLcosa5TUlqa/pVgGki5K12nbunVreUtvZjbBVfoy1zrgWJKphWcBd0k6ISK2A3MjYrOko4E7Jf02Ih7NbhwR1wLXQjIn9egW3cxsfMuzBrEZmJ15PStNy9oELIuIroh4DHiEJGAQEZvTv+tILnc/JceymplZH3kGiJXAsZLmS2oAzgf6Xo30I5LaA5KmkzQ5rZN0qKRJmfTXAA9hZmajJrcmpojolnQxcDtQC1wfEWskXQm0RcSydNmbJD0E9AB/HxHPSvoj4GuSekmC2OezVz8Vs2rVqmckrR9BkacDz4xg+2oxXo4DfCzVarwcy3g5DhjZscztb4Ei3HQPIKktIporXY6RGi/HAT6WajVejmW8HAfkdyy+k9rMzIpygDAzs6IcIPa5ttIFKJPxchzgY6lW4+VYxstxQE7H4j4IMzMryjUIMzMrygHCzMyKmtABQtKfp8OM90pqzqTPk9SRGW78mkqWsxT9HUu67OPpkOsPSzq7UmUcDklXSNqc+SzOrXSZhqKUIe/HCkmPS/pt+jm0Vbo8QyHpeklPS3owk3aYpJ9K+kP699BKlrFU/RxLLv8nEzpAAA8CbwPuKrLs0Yg4OX18YJTLNRxFj0XSApK72I8HFgFfSYdiH0v+LfNZ3FrpwpQqM+T9OcAC4IL08xjL/jj9HMba/QM3kHz/sy4FlkfEscDy9PVYcAMHHgvk8H8yoQNERPwuIh6udDnKYYBjWQwsjYg96XhXa0mGYrf8lTLkvY2CiLgLaO+TvBj4Zvr8m0DRUaOrTT/HkosJHSAGMV/SfZJ+Lul1lS7MCJQy7Hq1u1jSA2nVekw0A6TGw3ufFcBPJK2SdFGlC1MGR0TElvT5k8ARlSxMGZT9/2TcBwhJd0h6sMhjoF9yW4A5EXEK8HfAdyVNHZ0S92+Yx1L1BjmurwLHACeTfC7/UtHCTmyvjYhTSZrMPijpjEoXqFwiud5/LF/zn8v/SaXng8hdRLxhGNvsAfakz1dJepRkpNmKdswN51gobdj1iir1uCRdB9ySc3HKqerf+6HIDMH/tKQfkjShFeu/GyuekvTiiNgi6cXAmJ3RMiKeKjwv5//JuK9BDIekGYWO3HTComOBdZUt1bAtA86XNEnSfJJjubfCZSpZ+o9b8FaSzvixopQh78cESQdJmlJ4DryJsfVZFLMMuDB9fiHwnxUsy4jk9X8y7msQA5H0VuD/AjOA/5J0f0ScDZwBXCmpC+gFPhARo9IpNFz9HUs6xPpNJPNpdAMfjIieSpZ1iP5Z0skk1f/HgfdXtjil62/I+woXa7iOAH4oCZLzxncj4rbKFql0kpaQzD0zXdIm4FPA54GbJL0XWA/8ReVKWLp+jqUlj/8TD7VhZmZFuYnJzMyKcoAwM7OiHCDMzKwoBwgzMyvKAcLMzIpygDAzs6IcIMzMrCgHCLMBSHq/pKtHKa+mdHDIwl38R0paKunRdIC8WyW9VFKDpLskTegbXS1/DhBmAzsB+O0o5fUe4AcR0aPkluUfAq0RcUxEvBL4OMkIpJ0k8xf85SiVyyYoBwizgZ1InwAh6eWS7kxn7rpD0vQ0/bj0l/0Dkv5e0toh5vUO9o0H9MdAV0Tsnc0wIlZHxC/Slz9K1zfLjQOE2cBeQWbgM0mTgP8A/i4iTgZ+Cnwkbe75DvDhiDgROJohDJiWDuZ3dEQ8nsl31QCbPAi8agjHYTZkDhBm/ZA0G3guInZkkt8C3B0R96evHwJeRDLd6+qIuC+Tvjrdz9GSvi7p5vT1QZK+Kek6SYVawHRge6llSwdc7CyMsGqWBwcIs/4V639Y0CftBJJgcCJwfyb9FYXX6ZSj780sextwc0S8DzgvTesAGjPrrAFeOUj5JgG7Bz8Ms+FxgDDr3wH9DyQT/iyAvXOFvAv4FvAsyaRSpMMuv5O0BlHELPZNRdoDEBHbgFpJhSBxJzApO7WnpBML099KOhx4JiK6RnKAZgNxgDDr3wnARZIeTx8rgBuBmZJ+CywF3hMRz6bpzWn6e4HHI6K/SaY2kQQJ2P9/8CfAa2HvFJhvBd6QXua6BvgcydzJkHRi/1e5DtSsGM8HYVYGkg6OiOfT538PTIuIy9LXhwNXAW8E/h34EvBlkuahuyPiO+l6pwIfiYh3lZDfD4BLI+KRPI7HDBwgzMpC0idJphTtAn5JcpXTnmHs5z3ANwea9a8wfWlEfGu45TUrhQOEmZkV5T4IMzMrygHCzMyKcoAwM7OiHCDMzKwoBwgzMyvKAcLMzIpygDAzs6L+P1JfGNWbVmY+AAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([18]),)"
            ]
          },
          "metadata": {},
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fair_and_accurate2, stds_both2 = accurate_fair_combined_NONGLOBAL(accs1_G, eq_ops1_G)\n",
        "plot_graph(c_values, fair_and_accurate2, stds_both2, 'both', 'german')\n",
        "np.where(fair_and_accurate2 == np.min(fair_and_accurate2))"
      ],
      "metadata": {
        "id": "HjI_jI2xrxp2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 316
        },
        "outputId": "ee451acc-b5b8-4d3f-ed5a-e5fd71f9f9eb"
      },
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([22]),)"
            ]
          },
          "metadata": {},
          "execution_count": 57
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fair_and_accurate3, stds_both3 = accurate_fair_combined_NONGLOBAL(accs2, eq_ops2)\n",
        "plot_graph(c_values, fair_and_accurate3, stds_both3, 'both', 'adult')\n",
        "np.where(fair_and_accurate3 == np.min(fair_and_accurate3))\n"
      ],
      "metadata": {
        "id": "zRl7Wdp8SrnY",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 316
        },
        "outputId": "5df1366b-8c58-4be1-c5dc-59b7b0b25cd1"
      },
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([21]),)"
            ]
          },
          "metadata": {},
          "execution_count": 58
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "wb189YncQy9l"
      },
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "fair_and_accurate4, stds_both4 = accurate_fair_combined_NONGLOBAL(accs2_G, eq_ops2_G)\n",
        "plot_graph(c_values, fair_and_accurate4, stds_both4, 'both', 'german')\n",
        "np.where(fair_and_accurate4 == np.min(fair_and_accurate4))\n",
        "\n"
      ],
      "metadata": {
        "id": "p2W4HX06QzBw",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 316
        },
        "outputId": "78f40706-b6ae-44f3-a6b7-8b21c5a05ba3"
      },
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([21]),)"
            ]
          },
          "metadata": {},
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(np.log10(c_values), fair_and_accurate1, label='adult')\n",
        "plt.plot(np.log10(c_values), fair_and_accurate3, label='rw adult')\n",
        "plt.xlabel(\" $Log_{10}$(C)\")\n",
        "plt.ylabel('Z')\n",
        "plt.legend()\n",
        "plt.show()\n",
        "\n",
        "\n",
        "plt.plot(np.log10(c_values), fair_and_accurate2, label='german')\n",
        "plt.plot(np.log10(c_values), fair_and_accurate4, label='rw german')\n",
        "plt.xlabel(\" $Log_{10}$(C)\")\n",
        "plt.ylabel('Z')\n",
        "plt.legend()\n",
        "plt.show()\n",
        "\n",
        "#plt.plot(np.log10(c_values), fair_and_accurate1, label='adult')\n",
        "#plt.plot(np.log10(c_values), fair_and_accurate2, label='german')\n",
        "#plt.legend()\n",
        "#plt.show()\n",
        "\n",
        "#plt.plot(np.log10(c_values), fair_and_accurate3, label='rw adult')\n",
        "#plt.plot(np.log10(c_values), fair_and_accurate4, label='rw german')\n",
        "#plt.legend()\n",
        "#plt.show()"
      ],
      "metadata": {
        "id": "KbkPinCSTf1U",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 549
        },
        "outputId": "0785703d-9441-4de9-dacb-f17ae98a82c7"
      },
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(np.min(fair_and_accurate1))\n",
        "print(np.min(fair_and_accurate2))\n",
        "print(np.min(fair_and_accurate3))\n",
        "print(np.min(fair_and_accurate4))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4QCyQK-oEeQj",
        "outputId": "08434327-b424-44cc-b25a-6aab01c5e69f"
      },
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.7209350578668351\n",
            "0.0\n",
            "0.6677488252347898\n",
            "0.052228254640853025\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(c_values[18],c_values[22],c_values[21],c_values[21] )"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "lmhtMdCZOEck",
        "outputId": "9cf63cb5-a187-4ef8-a6e9-cb782555e3b2"
      },
      "execution_count": 80,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.00010481131341546875 0.029470517025518096 0.007196856730011528 0.007196856730011528\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fair_and_accurate1, stds_both1 = accurate_fair_combined(accs1, eq_ops1, accs2, eq_ops2)\n",
        "plot_graph(c_values, fair_and_accurate1, stds_both1, 'both', 'adult')"
      ],
      "metadata": {
        "id": "RK8EPVXfYswN",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "outputId": "4c747637-08b7-4706-e089-e52d7249abd2"
      },
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fair_and_accurate2, stds_both2 = accurate_fair_combined(accs1_G, eq_ops1_G, accs2_G, eq_ops2_G)\n",
        "plot_graph(c_values, fair_and_accurate1, stds_both1, 'both', 'adult')"
      ],
      "metadata": {
        "id": "DuLq8JllY6Pe",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "outputId": "593ee8fc-1e9c-4a2b-844a-83a3744a2b3d"
      },
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fair_and_accurate3, stds_both1 = accurate_fair_combined(accs2, eq_ops2, accs1, eq_ops1)\n",
        "plot_graph(c_values, fair_and_accurate1, stds_both1, 'both', 'adult')"
      ],
      "metadata": {
        "id": "rhLj32afYsy5",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "outputId": "d6603baa-5d83-4d16-e6b8-e76bf11eb026"
      },
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "fair_and_accurate4, stds_both1 = accurate_fair_combined(accs2_G, eq_ops2_G, accs1_G, eq_ops1_G)\n",
        "plot_graph(c_values, fair_and_accurate1, stds_both1, 'both', 'adult')"
      ],
      "metadata": {
        "id": "NqYnfXucYs0w",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "outputId": "22bfaac2-2095-4ec6-b46f-aee65e04d06b"
      },
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "pigC3cwdYs3G"
      },
      "execution_count": 65,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(np.log10(c_values), fair_and_accurate1, label='adult')\n",
        "plt.plot(np.log10(c_values), fair_and_accurate3, label='rw adult')\n",
        "plt.xlabel(\" $Log_{10}$(C)\")\n",
        "plt.ylabel('Z')\n",
        "plt.legend()\n",
        "plt.show()\n",
        "\n",
        "\n",
        "plt.plot(np.log10(c_values), fair_and_accurate2, label='german')\n",
        "plt.plot(np.log10(c_values), fair_and_accurate4, label='rw german')\n",
        "plt.xlabel(\" $Log_{10}$(C)\")\n",
        "plt.ylabel('Z')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "aFBmA8FKQzGf",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 551
        },
        "outputId": "cdc2d590-ae24-49f9-8ba8-e6297bc1a77f"
      },
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(np.min(fair_and_accurate1))\n",
        "print(np.min(fair_and_accurate2))\n",
        "print(np.min(fair_and_accurate3))\n",
        "print(np.min(fair_and_accurate4))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kGSz8_6yEsN7",
        "outputId": "16ffcafe-268c-4148-c700-3108c1d38514"
      },
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3.7129852060651354\n",
            "2.253368133495243\n",
            "3.468551114967607\n",
            "1.1006547847469057\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "def get_best(faa):\n",
        "  best_both = np.amin(faa)\n",
        "  print(best_both)\n",
        "  # what is the index of this?\n",
        "  index_bb = np.where(faa == best_both)\n",
        "  print(index_bb[0][0])\n",
        "  print(c_values[index_bb][0])\n",
        "  return c_values[index_bb]\n",
        "\n",
        "c1 = get_best(fair_and_accurate1)\n",
        "c2 = get_best(fair_and_accurate2)\n",
        "c3 = get_best(fair_and_accurate3)\n",
        "c4 = get_best(fair_and_accurate4)\n",
        "\n"
      ],
      "metadata": {
        "id": "8G5-O3c896DT",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "863cd864-a6ad-4668-c51c-0b6dd89d7479"
      },
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3.7129852060651354\n",
            "18\n",
            "0.00010481131341546875\n",
            "2.253368133495243\n",
            "22\n",
            "0.029470517025518096\n",
            "3.468551114967607\n",
            "18\n",
            "0.00010481131341546875\n",
            "1.1006547847469057\n",
            "21\n",
            "0.007196856730011528\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "dnHhI-DyB1QK"
      },
      "execution_count": 70,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "1jn_foCZB1Ul"
      },
      "execution_count": 71,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "ICOBa33iQzR_"
      },
      "execution_count": 71,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Plot normalised accuracy and fairness"
      ],
      "metadata": {
        "id": "d3j_ZWBAryu1"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# normalise data\n",
        "\n",
        "accs1_norm = np.array(accs1)\n",
        "accs1_norm = accs1_norm - np.amin(accs1_norm)\n",
        "accs1_norm = accs1_norm / (max(accs1)-min(accs1))\n",
        "\n",
        "\n",
        "\n",
        "eq_ops1_norm = np.array(eq_ops1)\n",
        "eq_ops1_norm = eq_ops1_norm - np.amin(eq_ops1_norm)\n",
        "eq_ops1_norm = eq_ops1_norm / (max(eq_ops1)-min(eq_ops1))\n"
      ],
      "metadata": {
        "id": "IcZp78RJr5t-"
      },
      "execution_count": 72,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# plot adult\n",
        "plt.scatter(np.log10(c_values), accs1_norm, label='Accuracy' )\n",
        "#plt.errorbar(np.log10(c_values), accs1_norm, yerr = stds1, fmt='.', ecolor = 'cyan', color='black', label='Adult')\n",
        "\n",
        "#plot german\n",
        "plt.scatter(np.log10(c_values), eq_ops1_norm, label='Fairness')\n",
        "#plt.errorbar(np.log10(c_values), accs1_G, yerr = stds1_G, fmt='.', ecolor = 'cyan', color='blue', label='German')\n",
        "\n",
        "plt.title('Effect of regularisation strength on normalised accuracy\\n and fairness: standard model')\n",
        "\n",
        "\n",
        "# C is the inverse of regularization strength\n",
        "plt.xlabel(\" $Log_{10}$(C)\")\n",
        "plt.ylabel(\"Normalised performance\")\n",
        "plt.legend(loc='best')\n",
        "plt.show()  "
      ],
      "metadata": {
        "id": "Z1QldL2cr2Ke",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "07d4a7c6-1eba-45b5-e29f-f35d380403b0"
      },
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# reweighed\n",
        "\n",
        "# normalise data\n",
        "\n",
        "accs2_norm = np.array(accs2)\n",
        "\n",
        "accs2_norm = accs2_norm - np.amin(accs2_norm)\n",
        "\n",
        "accs2_norm = accs2_norm / (max(accs2)-min(accs2))\n",
        "\n",
        "\n",
        "\n",
        "eq_ops2_norm = np.array(eq_ops1)\n",
        "eq_ops2_norm = eq_ops2_norm - np.amin(eq_ops2_norm)\n",
        "eq_ops2_norm = eq_ops2_norm / (max(eq_ops2)-min(eq_ops2))"
      ],
      "metadata": {
        "id": "gT9Mu8wCuLa6"
      },
      "execution_count": 74,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# plot adult\n",
        "plt.scatter(np.log10(c_values), accs2_norm, label='Accuracy' )\n",
        "#plt.errorbar(np.log10(c_values), accs1_norm, yerr = stds1, fmt='.', ecolor = 'cyan', color='black', label='Adult')\n",
        "\n",
        "#plot german\n",
        "plt.scatter(np.log10(c_values), eq_ops2_norm, label='Fairness')\n",
        "#plt.errorbar(np.log10(c_values), accs1_G, yerr = stds1_G, fmt='.', ecolor = 'cyan', color='blue', label='German')\n",
        "\n",
        "plt.title('Effect of regularisation strength on normalised accuracy\\n and fairness: fairness-based model')\n",
        "\n",
        "\n",
        "# C is the inverse of regularization strength\n",
        "plt.xlabel(\" $Log_{10}$(C)\")\n",
        "plt.ylabel(\"Normalised performance\")\n",
        "plt.legend(loc='best')\n",
        "plt.show() "
      ],
      "metadata": {
        "id": "eUr1mrMuuV7n",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "ebb152d7-f447-47f3-c888-7e5e590d877c"
      },
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "TyuGfFx2udZJ"
      },
      "execution_count": 75,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "RBvm1e5kudcg"
      },
      "execution_count": 75,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#German dataset\n",
        "\n",
        "# normalise data\n",
        "\n",
        "accs1_norm_G = np.array(accs1_G)\n",
        "\n",
        "accs1_norm_G = accs1_norm_G - np.amin(accs1_norm_G)\n",
        "\n",
        "accs1_norm_G = accs1_norm_G / (max(accs1_G)-min(accs1_G))\n",
        "\n",
        "\n",
        "\n",
        "eq_ops1_norm_G = np.array(eq_ops1_G)\n",
        "eq_ops1_norm_G = eq_ops1_norm_G - np.amin(eq_ops1_norm_G)\n",
        "eq_ops1_norm_G = eq_ops1_norm_G / (max(eq_ops1_G)-min(eq_ops1_G))\n"
      ],
      "metadata": {
        "id": "klWIYbq2udfh"
      },
      "execution_count": 76,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# plot adult\n",
        "plt.scatter(np.log10(c_values), accs1_norm_G, label='Accuracy' )\n",
        "#plt.errorbar(np.log10(c_values), accs1_norm, yerr = stds1, fmt='.', ecolor = 'cyan', color='black', label='Adult')\n",
        "\n",
        "#plot german\n",
        "plt.scatter(np.log10(c_values), eq_ops1_norm_G, label='Fairness')\n",
        "#plt.errorbar(np.log10(c_values), accs1_G, yerr = stds1_G, fmt='.', ecolor = 'cyan', color='blue', label='German')\n",
        "\n",
        "plt.title('Effect of regularisation strength on normalised accuracy\\n and fairness: standard model')\n",
        "\n",
        "\n",
        "# C is the inverse of regularization strength\n",
        "plt.xlabel(\" $Log_{10}$(C)\")\n",
        "plt.ylabel(\"Normalised performance\")\n",
        "plt.legend(loc='best')\n",
        "plt.show()  "
      ],
      "metadata": {
        "id": "3tSjGOD8uzEx",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "a5ae3766-f66c-4309-e7e7-183047ea9fbf"
      },
      "execution_count": 77,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# normalise data\n",
        "\n",
        "accs2_norm_G = np.array(accs2_G)\n",
        "\n",
        "accs2_norm_G = accs2_norm_G - np.amin(accs2_norm_G)\n",
        "\n",
        "accs2_norm_G = accs2_norm_G / (max(accs2_G)-min(accs2_G))\n",
        "\n",
        "\n",
        "\n",
        "eq_ops2_norm_G = np.array(eq_ops2_G)\n",
        "eq_ops2_norm_G = eq_ops2_norm_G - np.amin(eq_ops2_norm_G)\n",
        "eq_ops2_norm_G = eq_ops2_norm_G / (max(eq_ops2_G)-min(eq_ops2_G))\n"
      ],
      "metadata": {
        "id": "8SKuGoHmu7v-"
      },
      "execution_count": 78,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# plot adult\n",
        "plt.scatter(np.log10(c_values), accs2_norm_G, label='Accuracy' )\n",
        "#plt.errorbar(np.log10(c_values), accs1_norm, yerr = stds1, fmt='.', ecolor = 'cyan', color='black', label='Adult')\n",
        "\n",
        "#plot german\n",
        "plt.scatter(np.log10(c_values), eq_ops2_norm_G, label='Fairness')\n",
        "#plt.errorbar(np.log10(c_values), accs1_G, yerr = stds1_G, fmt='.', ecolor = 'cyan', color='blue', label='German')\n",
        "\n",
        "plt.title('Effect of regularisation strength on normalised accuracy\\n and fairness: fairness-based model')\n",
        "\n",
        "\n",
        "# C is the inverse of regularization strength\n",
        "plt.xlabel(\" $Log_{10}$(C)\")\n",
        "plt.ylabel(\"Normalised performance\")\n",
        "plt.legend(loc='best')\n",
        "plt.show() "
      ],
      "metadata": {
        "id": "uJqUXUISvHsI",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "outputId": "969d39c6-90cf-4f97-f6ae-b55c990af57b"
      },
      "execution_count": 79,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}