{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "WISDM.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/lisatwyw/data-gym/blob/master/WISDM.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LhHGvYR3KNKG",
        "colab_type": "text"
      },
      "source": [
        "# Overview #\n",
        "\n",
        "- This is a notebook prepared for you to try out classification tasks using time-series data acquired using accelerometers of iPhone's.\n",
        "\n",
        "- You'll get to:\n",
        " - train a model based on neural networks \n",
        " - learn about ```CoreML```\n",
        " \n",
        "- Post-demo exercise: try to replicate below using the [Actitracker dataset](http://www.cis.fordham.edu/wisdm/includes/datasets/latest/WISDM_at_latest.tar.gz)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FqM7vSoRHyg3",
        "colab_type": "text"
      },
      "source": [
        "# A) Mount Google drive #\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zKhYw_BxHyAa",
        "colab_type": "code",
        "outputId": "f70691c6-cf6f-4081-dd83-d0d8e7174b35",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')\n",
        "\n",
        "from pydrive.auth import GoogleAuth\n",
        "from pydrive.drive import GoogleDrive\n",
        "from google.colab import auth\n",
        "from oauth2client.client import GoogleCredentials"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rKZJs-RrIhKO",
        "colab_type": "text"
      },
      "source": [
        "# B) Create subfolder ``wisdm`` in your ```opensource_datasets``` folder #"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MRx2Rs5oIfSe",
        "colab_type": "code",
        "outputId": "12fa9bc6-22d8-4b07-de80-fb8c1dafd3d2",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "try:\n",
        "  ! mkdir '/content/drive/My Drive/Colab Notebooks/opensource_datasets/'\n",
        "except e as Exception:\n",
        "  pass  \n",
        "import os\n",
        "try:\n",
        "  ! mkdir '/content/drive/My Drive/Colab Notebooks/opensource_datasets/wisdm'\n",
        "except e as Exception:\n",
        "  pass  \n",
        "import os\n",
        "os.chdir('/content/drive/My Drive/Colab Notebooks/opensource_datasets/wisdm')"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "mkdir: cannot create directory ‘/content/drive/My Drive/Colab Notebooks/opensource_datasets/’: File exists\n",
            "mkdir: cannot create directory ‘/content/drive/My Drive/Colab Notebooks/opensource_datasets/wisdm’: File exists\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DXs3peqUI8QM",
        "colab_type": "text"
      },
      "source": [
        "# C) Download from website and unzip contents ##"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b4hQnzCbIgvF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "if os.path.isfile('wisdm.gz' )==False:\n",
        "  try:\n",
        "    ! wget -O wisdm.gz http://www.cis.fordham.edu/wisdm/includes/datasets/latest/WISDM_ar_latest.tar.gz\n",
        "    ! tar -xf wisdm.gz\n",
        "    print('done unzipping')\n",
        "  except:\n",
        "    pass\n",
        "\n",
        "  ! ls"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WuJS5m1SJvR1",
        "colab_type": "text"
      },
      "source": [
        "# D) Run the following demo code written by [Nils Ackermann](https://towardsdatascience.com/human-activity-recognition-har-tutorial-with-keras-and-core-ml-part-1-8c05e365dfa0?)# "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VlO0ei_RIxxY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "\n",
        "# https://towardsdatascience.com/human-activity-recognition-har-tutorial-with-keras-and-core-ml-part-1-8c05e365dfa0?\n",
        "\n",
        "def read_data(file_path):\n",
        "\n",
        "    column_names = ['user-id',\n",
        "                    'activity',\n",
        "                    'timestamp',\n",
        "                    'x-axis',\n",
        "                    'y-axis',\n",
        "                    'z-axis']\n",
        "    df = pd.read_csv(file_path,\n",
        "                     header=None,\n",
        "                     names=column_names)\n",
        "    # Last column has a \";\" character which must be removed ...\n",
        "    df['z-axis'].replace(regex=True,\n",
        "      inplace=True,\n",
        "      to_replace=r';',\n",
        "      value=r'')\n",
        "    # ... and then this column must be transformed to float explicitly\n",
        "    df['z-axis'] = df['z-axis'].apply(convert_to_float)\n",
        "    # This is very important otherwise the model will not fit and loss\n",
        "    # will show up as NAN\n",
        "    df.dropna(axis=0, how='any', inplace=True)\n",
        "\n",
        "    return df\n",
        "\n",
        "def convert_to_float(x):\n",
        "\n",
        "    try:\n",
        "        return np.float(x)\n",
        "    except:\n",
        "        return np.nan\n",
        " \n",
        "def show_basic_dataframe_info(dataframe):\n",
        "\n",
        "    # Shape and how many rows and columns\n",
        "    print('Number of columns in the dataframe: %i' % (dataframe.shape[1]))\n",
        "    print('Number of rows in the dataframe: %i\\n' % (dataframe.shape[0]))\n",
        "\n",
        "# Load data set containing all the data from csv\n",
        "df = read_data('WISDM_ar_v1.1/WISDM_ar_v1.1_raw.txt')\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kUEPILAbKbeS",
        "colab_type": "text"
      },
      "source": [
        "## Examine the data ##"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sxIPjT6cKWtS",
        "colab_type": "code",
        "outputId": "8c853c7a-267f-4368-b031-bd438f68c557",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        }
      },
      "source": [
        "df.head(5)"
      ],
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>user-id</th>\n",
              "      <th>activity</th>\n",
              "      <th>timestamp</th>\n",
              "      <th>x-axis</th>\n",
              "      <th>y-axis</th>\n",
              "      <th>z-axis</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>33</td>\n",
              "      <td>Jogging</td>\n",
              "      <td>49105962326000</td>\n",
              "      <td>-0.7</td>\n",
              "      <td>12.7</td>\n",
              "      <td>0.5</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>33</td>\n",
              "      <td>Jogging</td>\n",
              "      <td>49106062271000</td>\n",
              "      <td>5.0</td>\n",
              "      <td>11.3</td>\n",
              "      <td>1.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>33</td>\n",
              "      <td>Jogging</td>\n",
              "      <td>49106112167000</td>\n",
              "      <td>4.9</td>\n",
              "      <td>10.9</td>\n",
              "      <td>-0.1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>33</td>\n",
              "      <td>Jogging</td>\n",
              "      <td>49106222305000</td>\n",
              "      <td>-0.6</td>\n",
              "      <td>18.5</td>\n",
              "      <td>3.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>33</td>\n",
              "      <td>Jogging</td>\n",
              "      <td>49106332290000</td>\n",
              "      <td>-1.2</td>\n",
              "      <td>12.1</td>\n",
              "      <td>7.2</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "   user-id activity       timestamp  x-axis  y-axis  z-axis\n",
              "0       33  Jogging  49105962326000    -0.7    12.7     0.5\n",
              "1       33  Jogging  49106062271000     5.0    11.3     1.0\n",
              "2       33  Jogging  49106112167000     4.9    10.9    -0.1\n",
              "3       33  Jogging  49106222305000    -0.6    18.5     3.0\n",
              "4       33  Jogging  49106332290000    -1.2    12.1     7.2"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4DP3_8obMpUS",
        "colab_type": "text"
      },
      "source": [
        "## Prepare the output labels ##"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3tzzfcBqKag9",
        "colab_type": "code",
        "outputId": "6a68099e-8ab3-4a39-9c22-3d12d213f4ff",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import preprocessing\n",
        "\n",
        "\n",
        "# Set some standard parameters upfront\n",
        "pd.options.display.float_format = '{:.1f}'.format\n",
        "sns.set() # Default seaborn look and feel\n",
        "\n",
        "\n",
        "plt.style.use('ggplot')\n",
        "\n",
        "# Same labels will be reused throughout the program\n",
        "LABELS = ['Downstairs',\n",
        "          'Jogging',\n",
        "          'Sitting',\n",
        "          'Standing',\n",
        "          'Upstairs',\n",
        "          'Walking']\n",
        "\n",
        "\n",
        "# The number of steps within one time segment\n",
        "TIME_PERIODS = 80\n",
        "\n",
        "# The steps to take from one segment to the next; if this value is equal to TIME_PERIODS, then there is no overlap between the segments\n",
        "STEP_DISTANCE = 40\n",
        "\n",
        "\n",
        "# Define column name of the label vector\n",
        "LABEL = 'ActivityEncoded'\n",
        "# Transform the labels from String to Integer via LabelEncoder\n",
        "\n",
        "le = preprocessing.LabelEncoder()\n",
        "# Add a new column to the existing DataFrame with the encoded values\n",
        "df[LABEL] = le.fit_transform(df['activity'].values.ravel())\n",
        "\n",
        "num_classes = le.classes_.size\n",
        "print(list(le.classes_))"
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['Downstairs', 'Jogging', 'Sitting', 'Standing', 'Upstairs', 'Walking']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_VutlDPgMsbt",
        "colab_type": "text"
      },
      "source": [
        "## Split the dataset into training and test subsets ##"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Sp0jLBCWLKie",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pd.options.mode.chained_assignment = None  # default='warn'\n",
        "\n",
        "df_test = df[df['user-id'] > 28]\n",
        "df_train = df[df['user-id'] <= 28]\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QOJamwIzM2Ut",
        "colab_type": "text"
      },
      "source": [
        "## Normalize the input data ##"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ybe6nM2BM5Ke",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df_train['x-axis'] = df_train['x-axis'] / df_train['x-axis'].max()\n",
        "df_train['y-axis'] = df_train['y-axis'] / df_train['y-axis'].max()\n",
        "df_train['z-axis'] = df_train['z-axis'] / df_train['z-axis'].max()\n",
        "\n",
        "# Round numbers\n",
        "df_train = df_train.round({'x-axis': 4, 'y-axis': 4, 'z-axis': 4})"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TzaCXWHHLqMY",
        "colab_type": "code",
        "outputId": "e8445d85-ef9a-4194-fe59-28cb84e610d6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "import keras\n",
        "print('keras version ', keras.__version__)\n"
      ],
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "keras version  2.2.5\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YOlezCu2MVHB",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gDccK9tBNF0m",
        "colab_type": "text"
      },
      "source": [
        "## Prepare the training dataset ##"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aMfE8IclMmdD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from scipy import stats\n",
        "from keras.utils import np_utils\n",
        "\n",
        "\n",
        "def create_segments_and_labels(df, time_steps, step, label_name):\n",
        "\n",
        "    # x, y, z acceleration as features\n",
        "    N_FEATURES = 3\n",
        "    # Number of steps to advance in each iteration (for me, it should always\n",
        "    # be equal to the time_steps in order to have no overlap between segments)\n",
        "    # step = time_steps\n",
        "    segments = []\n",
        "    labels = []\n",
        "    for i in range(0, len(df) - time_steps, step):\n",
        "        xs = df['x-axis'].values[i: i + time_steps]\n",
        "        ys = df['y-axis'].values[i: i + time_steps]\n",
        "        zs = df['z-axis'].values[i: i + time_steps]\n",
        "        # Retrieve the most often used label in this segment\n",
        "        label = stats.mode(df[label_name][i: i + time_steps])[0][0]\n",
        "        segments.append([xs, ys, zs])\n",
        "        labels.append(label)\n",
        "\n",
        "    # Bring the segments into a better shape\n",
        "    reshaped_segments = np.asarray(segments, dtype= np.float32).reshape(-1, time_steps, N_FEATURES)\n",
        "    labels = np.asarray(labels)\n",
        "\n",
        "    return reshaped_segments, labels\n",
        "\n",
        "x_train, y_train = create_segments_and_labels(df_train,\n",
        "                                              TIME_PERIODS,\n",
        "                                              STEP_DISTANCE,\n",
        "                                              LABEL)\n",
        "\n",
        "# convert to float\n",
        "x_train = x_train.astype('float32')\n",
        "y_train = y_train.astype('float32')\n",
        "\n",
        "# convert output labels using one-hot-encoding\n",
        "y_train_hot = np_utils.to_categorical(y_train, num_classes)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jmIFj7ShNO65",
        "colab_type": "code",
        "outputId": "f2821931-a1b2-4dc9-d664-1a9f27e427c6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 86
        }
      },
      "source": [
        "num_time_periods, num_sensors = x_train.shape[1], x_train.shape[2]\n",
        "print( 'Input data:', x_train.shape, '\\n3 time-series and 80 timepoints; total of', y_train.shape, 'samples' )\n",
        "\n",
        "input_shape = (num_time_periods*num_sensors)\n",
        "\n",
        "\n",
        "x_train = x_train.reshape(x_train.shape[0], input_shape)\n",
        "\n",
        "print('After reshaping, x_train has shape', x_train.shape)\n",
        "print('input_shape:', input_shape)"
      ],
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Input data: (20868, 80, 3) \n",
            "3 time-series and 80 timepoints; total of (20868,) samples\n",
            "After reshaping, x_train has shape (20868, 240)\n",
            "input_shape: 240\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IRHcLJHAOqwJ",
        "colab_type": "text"
      },
      "source": [
        "## Construct the model ##"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GL4ztYe-N1ph",
        "colab_type": "code",
        "outputId": "ad6503f1-9156-47a7-cb1e-8f4b1dff1c01",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 382
        }
      },
      "source": [
        "import tensorflow as tf\n",
        "\n",
        "model_m = tf.keras.models.Sequential()\n",
        "# Remark: since coreml cannot accept vector shapes of complex shape like\n",
        "# [80,3] this workaround is used in order to reshape the vector internally\n",
        "# prior feeding it into the network\n",
        "model_m.add(tf.keras.layers.Reshape((TIME_PERIODS, 3), input_shape=(input_shape,)))\n",
        "model_m.add(tf.keras.layers.Dense(100, activation='relu'))\n",
        "model_m.add(tf.keras.layers.Dense(100, activation='relu'))\n",
        "\n",
        "model_m.add(tf.keras.layers.Dense(100, activation='relu'))\n",
        "model_m.add(tf.keras.layers.Flatten())\n",
        "model_m.add(tf.keras.layers.Dense(num_classes, activation='softmax'))\n",
        "print(model_m.summary())\n"
      ],
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "reshape_1 (Reshape)          (None, 80, 3)             0         \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 80, 100)           400       \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 80, 100)           10100     \n",
            "_________________________________________________________________\n",
            "dense_6 (Dense)              (None, 80, 100)           10100     \n",
            "_________________________________________________________________\n",
            "flatten_1 (Flatten)          (None, 8000)              0         \n",
            "_________________________________________________________________\n",
            "dense_7 (Dense)              (None, 6)                 48006     \n",
            "=================================================================\n",
            "Total params: 68,606\n",
            "Trainable params: 68,606\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xRQpFOMXOtL1",
        "colab_type": "text"
      },
      "source": [
        "## Train it! ##"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qFt6hd052JkS",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 642
        },
        "outputId": "abd5d818-d469-4999-9ced-5e8e9b75bca2"
      },
      "source": [
        "callbacks_list = [\n",
        "    tf.keras.callbacks.ModelCheckpoint(\n",
        "        filepath='best_model.{epoch:02d}-{val_loss:.2f}.h5',\n",
        "        monitor='val_loss', save_best_only=True),\n",
        "    tf.keras.callbacks.EarlyStopping(monitor='accuracy', patience=1)\n",
        "]\n",
        "\n",
        "model_m.compile(loss='categorical_crossentropy',\n",
        "                optimizer='adam', metrics=['accuracy'])\n",
        "\n",
        "# Hyper-parameters\n",
        "BATCH_SIZE = 400\n",
        "EPOCHS = 50\n",
        "\n",
        "# Enable validation to use ModelCheckpoint and EarlyStopping callbacks.\n",
        "history = model_m.fit(x_train,\n",
        "                      y_train_hot,\n",
        "                      batch_size=BATCH_SIZE,\n",
        "                      epochs=EPOCHS,\n",
        "                      callbacks=callbacks_list,\n",
        "                      validation_split=0.2,\n",
        "                      verbose=1)"
      ],
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "42/42 [==============================] - 7s 163ms/step - loss: 1.0735 - accuracy: 0.6414 - val_loss: 0.7456 - val_accuracy: 0.8146\n",
            "Epoch 2/50\n",
            "42/42 [==============================] - 6s 149ms/step - loss: 0.5890 - accuracy: 0.7921 - val_loss: 0.6748 - val_accuracy: 0.8023\n",
            "Epoch 3/50\n",
            "42/42 [==============================] - 6s 149ms/step - loss: 0.5184 - accuracy: 0.8117 - val_loss: 0.6935 - val_accuracy: 0.7911\n",
            "Epoch 4/50\n",
            "42/42 [==============================] - 6s 148ms/step - loss: 0.4905 - accuracy: 0.8214 - val_loss: 0.6859 - val_accuracy: 0.7748\n",
            "Epoch 5/50\n",
            "42/42 [==============================] - 6s 149ms/step - loss: 0.4700 - accuracy: 0.8277 - val_loss: 0.6886 - val_accuracy: 0.7777\n",
            "Epoch 6/50\n",
            "42/42 [==============================] - 6s 149ms/step - loss: 0.4455 - accuracy: 0.8389 - val_loss: 0.6566 - val_accuracy: 0.8052\n",
            "Epoch 7/50\n",
            "42/42 [==============================] - 6s 147ms/step - loss: 0.4338 - accuracy: 0.8420 - val_loss: 0.7045 - val_accuracy: 0.7590\n",
            "Epoch 8/50\n",
            "42/42 [==============================] - 6s 147ms/step - loss: 0.4165 - accuracy: 0.8480 - val_loss: 0.7055 - val_accuracy: 0.7702\n",
            "Epoch 9/50\n",
            "42/42 [==============================] - 6s 148ms/step - loss: 0.3936 - accuracy: 0.8573 - val_loss: 0.6924 - val_accuracy: 0.7683\n",
            "Epoch 10/50\n",
            "42/42 [==============================] - 6s 148ms/step - loss: 0.3802 - accuracy: 0.8618 - val_loss: 0.7079 - val_accuracy: 0.7676\n",
            "Epoch 11/50\n",
            "42/42 [==============================] - 6s 148ms/step - loss: 0.3662 - accuracy: 0.8662 - val_loss: 0.7503 - val_accuracy: 0.7288\n",
            "Epoch 12/50\n",
            "42/42 [==============================] - 6s 147ms/step - loss: 0.3546 - accuracy: 0.8679 - val_loss: 0.7465 - val_accuracy: 0.7547\n",
            "Epoch 13/50\n",
            "42/42 [==============================] - 6s 147ms/step - loss: 0.3432 - accuracy: 0.8730 - val_loss: 0.8359 - val_accuracy: 0.7015\n",
            "Epoch 14/50\n",
            "42/42 [==============================] - 6s 143ms/step - loss: 0.3325 - accuracy: 0.8777 - val_loss: 0.7773 - val_accuracy: 0.7487\n",
            "Epoch 15/50\n",
            "42/42 [==============================] - 6s 144ms/step - loss: 0.3282 - accuracy: 0.8779 - val_loss: 0.7938 - val_accuracy: 0.7319\n",
            "Epoch 16/50\n",
            "42/42 [==============================] - 6s 144ms/step - loss: 0.3209 - accuracy: 0.8807 - val_loss: 0.8080 - val_accuracy: 0.7183\n",
            "Epoch 17/50\n",
            "42/42 [==============================] - 6s 144ms/step - loss: 0.3029 - accuracy: 0.8881 - val_loss: 0.8184 - val_accuracy: 0.7417\n",
            "Epoch 18/50\n",
            "42/42 [==============================] - 6s 148ms/step - loss: 0.3041 - accuracy: 0.8863 - val_loss: 0.8976 - val_accuracy: 0.6950\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jJLDHSSuQIes",
        "colab_type": "text"
      },
      "source": [
        "## Examine the training progress and evaluate the trained model ##"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "b8XUbhKvQHdB",
        "colab_type": "code",
        "outputId": "c55a4ce7-7d72-4d15-c043-30c74ac44b45",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 528
        }
      },
      "source": [
        "plt.figure(figsize=(6, 4))\n",
        "plt.plot(history.history['accuracy'], 'r', label='Accuracy of training data')\n",
        "plt.plot(history.history['val_accuracy'], 'b', label='Accuracy of validation data')\n",
        "plt.plot(history.history['loss'], 'r--', label='Loss of training data')\n",
        "plt.plot(history.history['val_loss'], 'b--', label='Loss of validation data')\n",
        "plt.title('Model Accuracy and Loss')\n",
        "plt.ylabel('Accuracy and Loss')\n",
        "plt.xlabel('Training Epoch')\n",
        "plt.ylim(0)\n",
        "plt.legend()\n",
        "plt.show()\n",
        "\n",
        "# Print confusion matrix for training data\n",
        "y_pred_train = model_m.predict(x_train)\n",
        "# Take the class with the highest probability from the train predictions\n",
        "max_y_pred_train = np.argmax(y_pred_train, axis=1)\n",
        "\n",
        "\n",
        "from sklearn.metrics import classification_report\n",
        "print(classification_report(y_train, max_y_pred_train))\n"
      ],
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEdCAYAAAAb9oCRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOzdeXxU5dnw8d+c2WeSTJKZ7CsEkrDL\nFhZZZFFUFhUQKFXBhbpU1Nf2UdGqrbb49KlKbaVWXGtrVSiiIIpbRWQTkH03ECAbJJnsmSUzc877\nxyQDAwEmkA28v5/PMDNnmXNNSM51zr2qFEVREARBEIQQSe0dgCAIgnBpEYlDEARBaBaROARBEIRm\nEYlDEARBaBaROARBEIRmEYlDEARBaBaROIQ28+GHH9K9e/dm7fPXv/6Vq6++upUiEi7UY489xuzZ\ns9s7DKGdiMQh8Nhjj5GVlcX9999/xrqvvvqKrKysZp/w28OOHTvo1q0bU6ZMae9QBCArK4uPP/64\nvcMQWoFIHAIAiYmJrF69mrKysqDlH3zwAUlJSe0UVfO8//77/OxnPyM/P599+/a1dzgoioLH42nv\nMAShxYnEIQCQlpZGnz59+PDDDwPLioqKWL9+PZMnTz5j+2+//ZbJkyfTs2dPhgwZwm9/+1scDkdg\nvSzL/PnPf2bIkCH07duXhx56iOrq6jM+Z926dcyYMYPevXszfPhw5s2bR0VFRbPjr6mp4bPPPmP6\n9Olcd911vP/++2dsc+zYMR544AFycnLo06cPEydO5Jtvvgms3717N3feeSf9+vWjb9++TJ06lR07\ndgBNF5lt2bKFrKwsCgoKgJNFcRs3buTGG2+kV69erF+/nvz8fO6//36GDRsWOO5HH310Rnzvvvsu\n119/feBnOnfu3MCxx40bd8b28+bNY9asWWf9maxbt45bb72VnJwc+vfvzy233MLOnTuDtsnKyuLd\nd9/lf/7nf+jbty8jRozg1VdfDdqmsrKShx56iCuuuIKhQ4eyYMECWmLAiWXLlgW+74gRI1iwYAFe\nrzewfsuWLcyYMYO+ffvSt29fJk2axHfffRdY//e//50xY8bQs2dPBg8ezJ133onL5brouITzE4lD\nCJg2bRr/+c9/AieFJUuWMHjwYBITE4O2279/P/feey8DBgzg448/5n//939ZvXo1Tz/9dGCbf/7z\nn7z99ts88sgjfPjhh/To0YOXX3456HM2bNjAfffdx/jx41m+fDkLFy6koKCAuXPnNvvEtHz5cjp3\n7kxWVhaTJ09mxYoVQYmstLSUGTNmUF1dzd/+9jdWrFjBgw8+iCT5/wR+/PFHbrnlFiwWC//4xz9Y\ntmwZs2fPRpblZsUhyzLPP/88jz32GJ999hm9evXC4XAwePBgXn/9dVasWMG0adN4/PHH2bhxY2C/\nv/zlLzz//PPMnDmTFStW8PrrrweKB2+++Wby8/PZtGlTYPva2lpWrVrF9OnTzxqLw+HgZz/7Ge+/\n/z7vv/8+aWlp3HXXXWck5oULFzJw4EA+/vhj7r77bl588UU2bNgQWP/EE0+wZ88eXnnlFf7xj39Q\nWFjIl19+2ayfy+lWr17N448/zg033MCKFSt47LHH+Pe//x34HfF6vdx333306dOHZcuWsWzZMubO\nnYvRaATgiy++YNGiRTzxxBN88cUXvPXWWwwfPvyiYhKaQRF+8h599FFl1qxZisvlUnJycpQNGzYo\nXq9XGT58uPL5558rS5cuVbp16xbY/te//rUyZcqUoM/48ssvlaysLKWgoEBRFEUZPny48uKLLwZt\nM3fu3KDPueWWW5Q//elPQdsUFhYqmZmZyt69exVFUZS//OUvytixY8/7HSZNmqS88847gffjxo1T\nFi9eHHi/YMECZejQoUpdXV2T+//6179WJk6cqPh8vibXNxXH5s2blczMTCU/P19RFEVZunSpkpmZ\nqWzevPm88d5zzz3KE088oSiKotTV1Sm9evVSXn/99bNuf/fddyu/+tWvAu/fe+89ZdCgQYrb7T7v\nsRr5fD5lwIAByscffxxYlpmZqTz77LNB21177bXK888/ryiKohw5ckTJzMxU1q5dG1jvdruVYcOG\nKbNmzTrn8TIzM5WPPvqoyXU/+9nPlAceeCBo2dtvv6306tVLcbvdSmVlpZKZmals3Lixyf3feust\n5ZprrlHq6+vPGYPQOjTtnbiEjkOv1zNp0iSWLFlCXV0dPp+PUaNGsWLFiqDtcnNzGTx4cNCynJwc\nFEUhNzcXi8XCiRMn6Nu3b9A2/fr146uvvgq837VrF9u3b+fdd989I5YjR47QrVu3kOLesWMHhw4d\nYsKECYFlN910E++//z4333wzAHv27KFv376YTKYmP2PPnj0MHz48cAdyMXr16hX03ul0snDhQr75\n5htKS0vxeDzU19czaNAgwP/zdLvdXHnllWf9zBkzZvDAAw9QVVWFxWJh8eLF3HTTTeh0urPuk5+f\nz1/+8he2b9+O3W5HURScTidFRUVB22VnZwe9j42NDdR15ebmAgT9X+p0usCd1IXKzc3l+uuvD1qW\nk5OD2+0mPz+fjIwMbr75Zu68804GDx5MTk4OY8eOpXPnzgBcd911vPPOO4waNYphw4YxePBgxo4d\nS1hY2AXHJIROJA4hyPTp07npppsoLi5m8uTJaLXaVjuWLMvMmTOHG2644Yx1Npst5M95//338Xg8\nQSdeRVGQZZl9+/aFnIDORaVSnVF8dmp5fCO1Wo1erw9a9n//9398/fXXPPbYY3Tu3Bmj0cj//u//\nUltbG/LxR4wYgdVq5eOPP2bgwIHs2bOH559//pz73HPPPURFRfHUU0+RkJCAVqtl5syZZ1TYn/5/\n3NR3bQ+///3vue2221i3bh3r1q3jpZde4sknn2TGjBnExcWxatUqNm7cyMaNG3nllVd4/vnnWbJk\nCQkJCe0d+mVP1HEIQbp06UKvXr3YunVr4Gq9qW02b94ctGzTpk2oVCq6du1KWFgYcXFxbNu2LWib\nrVu3Br3v2bMnubm5pKWlnfEwm80hxdtYKf7UU0/x0UcfBR6NJ9jGSvIePXqwbdu2s14l9+jRgw0b\nNpy1TsNqtWK32/H5fIFle/fuDSnGLVu2MHHiRK6//nqys7NJSUnhyJEjgfUZGRno9XrWrVt31s+Q\nJImbb76ZJUuWsHjxYgYOHBi4+m5KRUUFubm5zJkzh+HDh9OlSxf0ej12uz2kmBt16dIFIOj/sr6+\nnl27djXrc5r63KZ+hwwGAykpKYFlmZmZ3H777bz++utMmTKFxYsXB9bpdDpGjBjBI488wooVK3C5\nXEF3tELrEYlDOMMbb7zBxo0bSU1NbXL9nXfeyd69e5k/fz6HDh1izZo1/P73v2fixImBivQ77riD\nd955h48++ogjR47w5ptvBlW4AjzwwAN8/fXXPPfcc+zbt49jx46xZs0aHn/88ZBbxyxfvhyVSsWU\nKVPIzMwMekycODFQST5z5kxkWea+++7jhx9+ID8/n2+++YZvv/0WgLvuuoujR4/y61//ml27dnHs\n2DE+++yzwAlz0KBBuFwu/vKXvwTWNVXE1pROnTrx9ddfs3PnTnJzc3nyyScpKSkJrDebzdx+++28\n/PLLvPvuu+Tl5bF///4zWjdNnTqVw4cPs2TJEqZNm3bOY1osFqKjo1myZAl5eXls27aNhx9+GIPB\nEFLMjdLS0hg9ejTPPPMMGzduJDc3l9/85jfU1dWFtH9xcTH79u0Letjtdu6+++5ABXdeXh6ffvop\nL7/8Mrfffjs6nY6jR4/ypz/9iS1btlBYWMi2bdv44YcfyMjIAAgk0P3791NYWMjy5cupq6sLJDqh\ndYnEIZzBaDQSGRl51vXZ2dm88sorbNmyhRtuuIFHHnmEkSNH8rvf/S6wzW233catt97Kc889x403\n3sj27dv55S9/GfQ5gwcP5h//+AcHDhxg5syZTJo0ieeeew6z2YxGE1op6uLFixk1alSTJ8Srr74a\nt9vNypUriY2N5d///jdms5lf/OIXTJgwgQULFgS2zcrK4p///Cfl5eXceuut3HDDDbz11luo1WoA\nOnfuzLPPPssnn3zChAkTWLp0KQ8//HBIMc6bN4+kpCRuu+02Zs+eTVxc3BnNax966CEeeugh3nnn\nHSZOnMgdd9zBnj17graJjY3lqquuwmw2c+21157zmJIk8dJLL3Hs2DEmTZrEY489xqxZs4iJiQkp\n5lPNnz+f7Oxs7rnnHm655Rbi4uJC7s2/YMECbrzxxqDH0qVLGTlyJPPnz2fZsmVMnDiR5557jpkz\nZwY6oRqNRo4ePcrDDz/MuHHjmDt3Ln379uWpp54C/Inxww8/5NZbb+X666/n7bff5plnnmHIkCHN\n/n5C86mUjlCYKQhCSKZOnUq/fv14/PHH2zsU4SdM3HEIwiWgvLycDz/8kL1793Lrrbe2dzjCT5xo\nVSUIl4AhQ4ZgsVh44okngiqPBaE9iKIqQRAEoVlEUZUgCILQLCJxCIIgCM0iEocgCILQLJdd5fjp\n4/CEKjEx8YL3bS8i5rZxqcV8qcULIua2craYTx8B+3zEHYcgCILQLCJxCIIgCM0iEocgCILQLCJx\nCIIgCM0iEocgCILQLCJxAO4jRznwwgJ8zZhYRxAE4adKJA5AdjopW7OW+vyC9g5FEAShwxOJA9Cl\nJANQX1DYzpEIgiB0fCJxAOqwMHRWK25xxyEIgnBeInE0MHdOp75AJA5BEITzEYmjQXjXrqg0WhRZ\nbu9QBEEQOrTLbqyqC5Uy/WbUw69s7zAEQRA6PHHHIQiCIDSLSBynOPH3RZQv+7i9wxAEQejQROI4\nhexw4voxt73DEARB6NBE4jiFLjmZ+sIiUUEuCIJwDm1SOf7OO+/w/fffU1payvPPP09qauoZ28iy\nzJtvvsmOHTsAuPHGGxkzZkxbhBegS0lCqa/HU1KKLj6uTY8tCIJwqWiTO46cnBx+97vfERMTc9Zt\nvvvuO06cOMFLL73EH/7wB5YsWUJJSUlbhBdwsge56M8hCIJwNm2SOLKzs7HZbOfcZv369YwZMwZJ\nkoiIiGDgwIFs3LixLcIL0MXHY+zZA8lobNPjCoIgXEo6TD+OsrKyoORis9koKytr0xhUGg3xv7yn\nTY8pCIJwqekwiaOlNHfS9ab29blcSHo9KpWqpcJqNRfzfduLiLn1XWrxgoi5rbREzB0mcTTeYXTp\n0gXw34Gcq07kbIqKii7o+ImJiRQVFVGzYSNl//w3KfOfRRNpuaDPaiuNMV9KRMyt71KLF0TMbeVs\nMTc3mXSY5rhDhgzh66+/RpZlqqur2bx5M4MHD27zODQ2GyiKqCAXBEE4iza543jzzTfZtGkTlZWV\nPPvss4SHh/Piiy/y3HPPMW3aNDIyMhgxYgQ//vgjDz74IABTp04lNja2LcILok9OAqA+vwBTzx5t\nfnxBEISOrk0Sxx133MEdd9xxxvJ58+YFXkuSxJw5c9oinHOSjEY0NpuY1EkQBOEsOkxRVUeiS0nG\nLYqqBEEQmtRhKsc7kvAhg/Da7SiKckm0rBIEQWhLInE0wdSrZ3uHIAiC0GGJxNEERVHw2u2ggDbm\n3D3eBUEQfmpEHcdZFP3xBSpXfd7eYQiCIHQ4InE0QaVSoUtOoj5ftKwSBEE4nUgcZ6FPSaa+uBjF\n52vvUARBEDoUkTjOQpecDF4vnuMn2jsUQRCEDkUkjrNonJvDnZ/fzpEIgiB0LKJV1Vlo42KJvesO\n9F0y2jsUQRCEDkUkjrNQSRLm/n3bOwxBEIQORxRVnYOnpITqb79DUZT2DkUQBKHDEInjHJz7D2J/\nfzHe8or2DkUQBKHDEInjHBoryMXcHIIgCCeJxHEOuqREUKmozxeJQxCEjim/xIEst21xukgc5yDp\ndGjjYsUdhyAILU7x+fBVV+OtqsJXW4vsdCLX16P4fCHXq36/z85z7+7l883FrRxtMNGq6jx0ycm4\nD+e1dxiCIHRwcn09vppa5NpafDU1+OrqkGtq8dX6H/7lja/rkB2Oc3+gJKFSq0GtRtXwIPAssVbf\nhdW6LqTLlWSt34Sn01S0bTRrqkgc5xE95SYkg769wxCEnyxFllHcbhSP99Sl/n+DLsyVJl8GL1f8\nn+fxoHi8/mev1//weE4u956y/oxt/M8VsoyjrMyfCGpqUTyepr+AJKEOD0MKC0MdFoY+JeXke7MZ\nJMk/tJHP57/bOOU1PvnkMtn/7PH6WF4bzzZPNH3UdiZp89CoJaDt5g4SieM8NJGW9g5BEC45is+H\n7HIhu1woThey2+0vinG7UVzuk+tcbmS3ixqVhKOiomG5f1njdkp9fXt/Hf+VvlaLSqM5+WyJQB0e\nji4h3p8EwsNQh4U3JAhzIDlIRmOLTghXUOpg97/3MWFIPBOGDGiXyeZE4jgPRZapWPEp+rQUzFf0\nae9wBKHFKLLccGUrn3JFK5+84vV6/SdypzPwrLhcyM6GR1PrGtaHerJXabWoDAY8ZjOyRoNkNKC2\nRKA1xCAZDKj0eiSDwf9aozntovqUN6eePFVNrT/lpVqNSqNFpdWcTATaxvenPTcmCenM6uDExESK\niopC+p4twen2YdSrSY4x8cztPYmOaL+SEJE4zkMlSdRu2IivolwkDqHV+a/U3Shu/1W64nY3XIE3\nvD7PMsXt5oSi4HG5gpNAIEmcfM2FdmxVqfwn8oYTumQ0oDabkWw2JGPDcqMxsE7S+59PJgE9Kn3D\ns1oNtP1J+FJz5HgdCz/6kcnDkxnSw9auSQNE4giJLiUZt2iSKzRBkWV/Raer4UTf8Bx0gne5gk7s\nstuF7K4PbNtYXCO768HrPf9BG6j0OiS93n9Cbnw2mTBaIlB5PCCdUqnaUNHaWLEaqGgNqoCVgvfR\nqIMTQENCUOl0TV6BC61je24Fb6w8TLhJQ1q8ub3DAUTiCIkuORnn3n3IHg+SVtve4QhtQFEU5Lo6\nfJVVVBw/Qc3hw3grq/BVNT4amlFWVYMsn/8DNRokvb7halsfuGLXWiwnlzUuP3W7xoRg0AcniXOc\nvMXV++Xj663HWfJNPmnxZu67sSsWc8c4/4jEEQJ9ShLIMp6iYvRpqe0djnCBFFn2l8k7HMgOJ766\nOnzV1f5EUNmYCKrwVVbhra4OXP2fOg+kZDajjrSgsVgwJiSgtljQWMJRGY3+E79e7y+qOf3krxF/\nakLz5BXXsvibfK7oEsmd13dGp1W3d0gB4rc5BLqUZFQGA76qqvYO5SdN8flQ6utRPB5kpwufwxFI\nAvIpr08uP7nO53CguFxn/WzJaERtsaC2RGDomtHw2oIm0kJcRgYVXg/qiAhxxym0OkVRUKlUdEoI\nY+7kTLqnRSBJbd9y6lxE4giBxmYj7YU/inLdZlIUBbmmFm95OV57Od6KcjxqDTXlFcie+lPaydej\n1De0la9veO/xIgfa1fsfoRQJqbRaJJMRyWRCMplQR0WiS0oIvPc/jKgb11siUFssSDrdWT8zIjGR\nWlH0I7SBqjoPi1bkMnl4MhlJ4fTs1DG7A4jEEQKVShXc3E8A/EU/vqoqvOUV/sTQmCDKT74+vVNU\nuUrlb96o06LS6k42eWx4SEYDKku4vzmkThu0TqXVIgW2MwYliMaEIO4IOg5ZUfhwTQFHjtcx7aoU\nUuM6RsVuR1VU5uTlZQepcXhxuH3tHc45icQRotpNW6j+dg0Jv3rosr/zUGTZ30qo7mSRj7eysiEp\nVOC12/3PFRX+Zp2nkMLC0ERHoU1IwNijO5roaDRWKxprFJroaJIzMigubttxdYT28f7XR/l2Ryl6\nrcRz7+5l7IB4Jg5J7FBl9R3F/mPV/H15LlqNxK9nZJPWwZNsSInjk08+oWfPnqSnp3Pw4EEWLFiA\nJEk8+OCDZGZmtnaMHYLi9eA+nIe3rKzNxoO5WIrXi69hTBy5rg6fw4nsqAskBN8picFXV9dQH+B/\nPlsbf7UlAo3Vij49FXO/K/xJIToKjTUaTXQ0kv7c7cvbo5er0D7S4sxcN0jD1QPiWfptPut2lXH1\ngHiROE6TV1zLS0sPEh9l4P7JXbG2cx+NUISUOFauXMno0aMBeO+995gwYQJGo5G3336b+fPnt2qA\nHUXj3Bzu/IJ2TxyKoiA7nTiOHcNx8Ed/K6DKSnyVlf4mow3Pck3N2T9EpWoo6jH7y/zNZrQxtob3\nJtRmE5L5ZDGQxmJBExWJShQFCefgkxUKSx2kxpm5sldMYPlt4zoxebiHMJMWWVFYtamYkX1iMRt+\nWoUePlmhqraeipp6Kms99M+KJi3OzIQhiYzuG4tRf2n8PEKK0uFwYDKZcDqdHDlyhCeffBJJknjn\nnXdCPlBRURELFy6ktraWsLAw7r//fhISEoK2qaqq4m9/+xt2ux2fz0ePHj24/fbbUatb9wqlyO7k\nnS83o5W82Cx6YiL1xFj0WC16DDr/sXXx8SBJ/rk5+vdr1Xh8DgeeEyVnJAL/cyW+yiqU+nqOnbaf\nZDajibSgjoxEl5qCJjISdXg4ktnk79nbWBdgNvn7C1zmRW5C2/L5ZN749DA7D1fyu9t7nXHlHGby\nX3TkFdexYl0h32w9wfTRafTPjGqPcFucrCjUOryU1/gTQ3mNm2G9YtBr1azeXsKq74uorPME3cz/\nOT0Co17D+MGJ7Rf4BQgpcVitVg4cOEB+fj7dunVDkiQcDgdSM048r732GuPGjWPEiBGsWbOGRYsW\n8fTTTwdts2zZMpKSkpg3bx5er5ennnqK77//nqFDhzbvWzWTLCucqHBQWFKD87RKqXCTxp9MLHqG\nWGy49uVRNqCGGIseS5gW6SKLXhSvl/rCItx5R3AfOYr7yBE8J0qCN1Kr0VgsqCMt6JOTUffsgSYq\nElt6J2oUGXVkJOpIi6gYFtqN1yfz2ieH2J5bydSRKecsbslIDOPxW3rwzhd5vPbJIb7vHMn/+3l0\nG0Z7YRRFocbp5US5i+PlLtzbK+nX2Uh0hJ4Ne8r415dH8PqCi3izUyNIspmICtOSnRZBVLiOqHAd\n0eF6osN1l2yxXUiJ45ZbbuHFF19Eo9Hwq1/9CoCtW7fSpUuXkA5SVVVFXl4eTz75JADDhg3jzTff\npLq6moiIiKBtXS4Xsizj9Xrxer1ER7f+L1RyjIkFD42kqKiIOpeX0ko3ZVVuyirdlFb5Xx8ursNA\nHPqqer78YD8AGrUKa4T/DsVm0QcSTHSEDqtFj0mvDirTVxQFb1kZ7ryjgSRRn1+A0tDRTAoPx5Ce\nRtigHHRJiaijItFYIpHCzE3eHcQkJuLp4M1EfbLCocIadh6qZN+xarqkFDOqTyTx0cb2Dk1oIR6v\nzKsrctl1uIrpo1IZ3S/uvPukxJp4bGZ3/rv1BB+vK+TZN7/nkeldO0QdmM8nU1rp5niFi+QYEzaL\nngPHqvn7ilwcrpMXlpKkIjk6k+gIPYk2I6P7xjUkBR3REf4EEWb0n2L7dImiT5fL484KQkwc/fr1\n49VXXw1aNnjwYAYPHhzSQex2O9HR0YE7FEmSiIqKoqysLChxTJ06lRdeeIG7774bl8vFtddeS3Z2\ndqjfpUWYDRrM8RrSmxgTxufrSXlNPd0bEkvpKcklt7AWV33w3YpBKxGlV7AoLiKclYRXniC8rhyL\nt45I3FhS4gi/agSG9HR06WlooqM6xB/OxXK4vOw5UsXOQ5XsPlKFw+VDLanonBjGxt3FfLutgJxs\nK+OHJBIXZWjvcIWL9N3OUnYdrmLm2DRG9gm9/k8tqbh6QDxXdInEGBaFSuWk3uPDXl1PgrX1Lyxq\nnV5QFMJMWipq6nnv66McL3dRWuUOTMXa+J2sFj0DsqKJjzYQF2UkPtpAj6x0Thz3txBMizN3+JZQ\nLSmkxFFQUEBYWBiRkZG4XC6WL1+OSqVi0qRJaFpwKIUNGzaQmprKk08+icvlYv78+WzcuDHkBAX+\ncXouVCj7puDvwaw6rd5FURQqSyvZ9+U6CvOKOH68EnulTJUmDLvWTJ7OijsiDk65wdLr1MQ6jcQe\nNxHr9hEb7SA2ykhstIkEqxlL2PlbV1zM921JxWV1bNp7nE17jrPnsB2frBBh1jG4ZyI53ePpmxWD\nyaClqtbNh9/ksnJ9Hpv327mqfwozrs4iwdax/+g6ys85VG0Z78z4BPp0S6FXhu2C9j8ZahT/+mwf\nS7/JZdqYrkwd0xWtpmWKcrw+mVUbjnCkuJpjx2soLK2luq6em8d05bbruxPp8lBRd5iMlChG9g8n\nKSaM5NgwUuLCMeo1JCZC726dmoj90vq9gJaJOaSz/ksvvcT/+3//j8jISN555x2Ki4vRarUsWrSI\nuXPnnnd/q9VKeXk5siwjSRKyLFNRUYHNFvyLtmrVKu69914kScJkMjFgwAB2797drMRxoYO7hTIw\nnFxfT8FTvyNi1Egix10TtE5RFI7/+a8oB38kJTKSLump6NPT0aenoU9LRTIYcLi82KvrsVe7Ka+u\np6zhubS8lv1HyqlzBY+MGmHWkmQzkmwzkhRjIslmJMFqRKuRQo65tciywuGiWnYcrmTXoUqKy/3D\neSRYDYztH0fvjEg6J4QFhkqoLC+lsiHmcf2jGJIdxuebi/l2ewGrf8hncHcb1w9OICay492BXGqD\nBrZFvK56H//+6ig3DEvCGqHHarzwvz04GfOAribyCiL59xcH+OaHo9x6TScyEsNC+gyfrHCiwkVh\nqYOCUieFZQ4SrEamjEhBURT+sXIvagkSbSb6ZFiIizLQNUEbiPuJn59euuGkwu6k4jwxX0rOFnNz\nk0lIiaOkpITExEQURWHTpk28+OKL6HQ67r///pAOYrFYSE9PZ+3atYwYMYK1a9fSqVOnM+o3YmJi\n2L59O126dMHr9bJr1y4GDRrUrC/UmiSdDpVGS31+4Rnr6jb/gOvgj1in30zEVSOa3N9k0GAyaEiJ\nNTW53lXvo7whsZyocAV++VfvKMHj9d86SyqIizKQaDPSLaOGCJ2XpBgT1ghdqxdzOd0+9h6pYufh\nSnYdrqLO5UWSVGQmhzO8Twy9O0eGfOKPMGu5+apUrhmYwOebivl2Rwkb99kZ2sPKdYMSsVk6flv2\nnyqn28dfPzxIXnEt/bpGtRBE9r0AACAASURBVGi/gwiTljvHZzCou5V3vzzKn97bx/TRqYzqG1xv\nUu3wUFjqwOH20T/TXw86/197KCh1Av76h4RoA6mx/jtZlUrFH+7qjdmgviyKg9tbSIlDp9PhdDop\nKCjAZrMRERGBz+fDc7Y5dpswZ84cFi5cyNKlSzGbzYGk89xzzzFt2jQyMjKYPXs2r732Gr/61a+Q\nZZkePXowZsyYC/tmrUSXkuRvknsK2emkfOkydGmphI8YdsGfbdCpSbQZSbQZ6XXq58sKJZUuCkud\nFJQ5KCx1cvSEgx8O7j9lX4kkm/+upPHuJMlmxGTQICsKXq+Cxyfj9cp4fAoer+x/+JSGZTIeb8Ny\nnxzY3uOVqffK5BXVcbCgBp+sYDao6dkpkt4ZkfRoaE54oSxmLdNGpXLNwHhWbSrmu52lrN9j58qe\nNq4flNDuE9YIwRwuLy8tPcixEgdzJmRwRdfWqfDt2SmSp2eHs3xdIdmp/gvMNTtL2HqwgsJSB9UO\n/915hEkTSBxXD4gH/I1d4qMNDfNwn9RYUS1cvJB+kldeeSXPPPMMTqeTa6+9FoC8vDxim9ERLikp\nqcnOgvPmzQu8jo+PD7S86qh0yck4duxCdrmRDP6TWsWKlfhqaoi77+5W6RshSSrio43ERxvpn3Wy\nlVlkdAw/7M6jsCGZFJQ62HKgnDU7SwPbaNSqM5oIXoj4aANj+jUUQSWGoW7h0Tojw3TMGJ3GuIEJ\nrNpUzNpdpazfXcawXjauG5RIVPjZByG8UG6PD0XxV9Kq1aqLblp9uatzevnzfw5QWObknokZrd5K\nyKBTM23UyWkM7FX11Lm89OwUSVKMkeSGC6RGg7tfWB2L0HwhJY7Zs2ezY8cO1Go1PXv2BPy3frNm\nzWrV4DoifUoyKAr1hYUYMjrjzi+gevUawocPa/O5OkwGLRmJYUFlwIqiUFFTT2GZk8JSJ3VuLzqN\nhFYtodGo0Da81ja+1kho1f7XmlOXN7zWqCU0alWb3d5Hhev42Zg0xg2M57NNxazdVca63WUM6xXD\ntTkJIScQRVGoc/kor3YH6pLsVe6gOqbTB5KTVKBWq1BLUiCZaCQVet1eFMWHRq1CkvzL1Gr/Nhq1\nCrWkwj8OZsMzIDW8lhrXcXIbqWHMzMD2Dc8GrZoB2dEktkGLoguhUoFWI3HvDV3o1TmyzY9/0/Bk\nbhqe3ObHFc4U8r1bnz59KCsr4+DBg0RHR5ORkdGacXVYurRUIkZfhWQ2ocgy9vc+QDKbiZo0ob1D\nA/wnoegIPdER+nb5424p0RF6fj42nWtzEvh0YzFrdpaydlcpI3r7E0iEWUuNw4u9+mQysFf5e+va\nq/zv3Z7gYdj1WgmrRY81QkeXpHCiwnVIKvDKCj6fgu+UZ2/je1lBpzNQW1cXvLyhuM9Vr+D1+Y8j\nK/6EpSg0PJTTlikoZ2znf5YVqPf4WLmxiOzUCEZdEUvvjMgOMQ9DtcODQavGZNDwPzOyRR2BEFri\nqKio4M9//jM//vgjYWFh1NTUkJmZyYMPPtgmHfQ6Eo3FgvXmKQDUrNuAO+8Itlm3oDY3XeEtXBxr\nhJ5br0nnukEJfLqxiNXbS1izsxSVSoXHG5wYTHp1oENmdloE1gidv8VPw7PpAitG26r1TI3Dw7pd\nZazeUcIry3OJDtcx8opYhvWKabfy+crael5cfIAkm5G7J3URSUMAQkwcr732GmlpacybNw+DwYDL\n5eK9997jtdde49FHH23tGDscxevFfayA8mUfo++SQdignPYO6bJns+i5bVwnrhuUwOrtJagAq8V/\nZ+VPDLpLZoC4swk3abl2UAJXD4xn56FKvtl2gmXfFfDJhkIGZlkZ1Tc25Dktahz+his6rb84MpT6\nm3qPzI+FNezJq2LPkSrqnP4K6HqvzK3XpF/w9xIuPyH9pR04cICHH3440NnPYDBwyy23cM8997Rq\ncG3JJ4degVzx8SdUff1fAGwzpomrsDYUE2ng5qsu73nf1ZKKvl2j6Ns1iqIyJ99sP8HGPXbW7ykj\nIzGMUX1j6ds1KqjVkMcrs/9YdaB48q3P8thz5ORUxzqNRFKMkcdmdgfg3S+PUFzuBFS43D5qnB6q\nHV5kWUGrUREbacBZ78PrU4iJ1HPxzSsuXLHdyacbiyircpMUYyI5xkRyQ+V44yCkQtsKKXGYzWYK\nCgpIT08PLCsqKsJkujyKZ2RFYf5bm7BFqJg4JOm85coqvQ4UBfNg/5hSgtBaEm1Gfj42nZuGJbN+\nTxnfbi/h9ZWHiTBrGdE7hgFZUWzLreS/W09Q4/DyzO29SEyE0f1i6dXZQr1Hxu2Rqff4MBk1uOp9\nHMiv4VBRLScqXIEWd5LK3zBh5th0MpPD+cO/9mDQqRnbP45vd5Tywgf76Z8ZxdSRKW3WRLq00sUn\nG4r4fp8dnUYiJdbEDwfK+e6UVoM2iz6QRBoTis2i7/AXc7KsUFDq4HBxHenx5iaHOOrIQkockyZN\n4tlnn2X06NHExMRQWlrK6tWrmT59emvH1yZkWcESpuPTjcfIK6rjzvGdCTc1PdKs4vNR+8NWAAwZ\nndsyTOEnzGTQMLZ/PKP7xbH3SBVfbjnOJxuK+GSDv+6lU4KZu67vTGyU/6Tes5P/zkNRFIrszobi\np2pWbijC61PQayW6p1no2clC9/SIMzpuzvt5dzRqf6u6Mf3i+HzzcT7ffJydh6sYNzCecQNbb0Km\n8mo3KzcWs353KWq1xNX9/ccLM2kDrQbzS50NPcT9vcR35FYG7ooa+zQlx5hIjjWSEmMi0WZE344j\n0SqKQrHdxf5j1RzIr+ZgQU3QgIl9MiKZODTprJ2DOxqVopxlqrfT7N69m7Vr11JRUUFUVBRXXnkl\nvXr1Ov+ObexihhxZ/PkO3vv6KOFGDb+Y2IXOTQx1UPXNt5Qv/g+o1USMGol1yk0XG/IFu5yGPOjI\nOlLMHq+MViNR6/TyxOs7sITpqKypx+2RSY01cVXfWMYOyWLdD7nsOVLFnrwqKmr99R2JNiM90v3J\nIiMxLDB0TajKq93859t8fjhYQXS4jikjU+if2TIDcyYmJrLv4BE+a+gECjC8dwzX5SRgCTt/E2y3\nx0dRmZP8UgcFJf4+TYVlDlz1/gYUKiAmSk9yjIlEq5G4KANx0QbiogwXXNx1rt8LRVEoqXRzIL+a\nA8dqOJBfTU1Dp0WbRU9WSjhZqRGkx5vZcqCcr7Ycx+H20a9rFBOHJpFoa50m2S015EjIieN0Xq+X\nZ599lt/97ncXsnurudixqo6dqOPVFYdQFIVn7ugVVI7sraqm4LfPYuiUjs/hQDIYSHjo/GN1tZaO\ndEILlYj5wuQW1PD55uNUOzw8NrMbKpUKp9uHUa/GVe/j+312Vm8rocjuDOxj0KnpnhZBj3QLPTpZ\nWqwT5cGCGj7471EKSp10TQ5n+qjUi7pSrnV4WLevlhVrD+PzyQztGcP4wRc/aoCiKNir6ykocfgT\nSsPdib3KHVRnYzFriYsyEBtlIC5KT1y0kbgo/zQJp/c+P9Xpvxfl1W72NySJA8eqAwk7MkxLVkoE\nWanhZKVENDmcjsPl5asfTvD11uO462UGZEczYUhii08/0KZjVTVFURT2799//g0vMalxZh6/pTvl\n1fVo1FJD230ZvVZN+YfLULxerDNuxlNmR9WCIwMLwulkRWHnoUq+2HycQ0W1hBk1jOobiywrqNUq\njHr/lbJBp2Zkn1hG9I7hYEENZbUSseEKnRPMqM9x4rtQmcnhPHFLD9buKuWjtYX84V97GN4rhhuu\nTArM8heKxpPlVz8cp94rM6iblfGDE4ltoaH2VSpVYJ6cU4dG8XhlSipdnKhwUVLuHxfuRIWL7bkV\n/qHWG0gq/91BXHRjUjn5iAzTUlHtYtM+uz9R5NdQWukG/EObNN5RZKdEEBt1/joXk0HDpCuTGN0v\nji+3FPPfrSVsOVDOoG5WJgxJ7HCDf4ozXxPMBk1gLuTl6wrZeaiSWb01KJu2EHn9tWhjY9t93nHh\n8rd5fzlvfnoYa4SOGaNTubKn7Zz1CiqViqyUCEa1wR2SJKkY0SeW/lnRfLKhiNXbTrDlQDkThiZy\nVZ/YcyYsV72P/247wZebjzcMUhjFHTdcgUaubdWYG2k1jeO6nXmXVOf0BpLKiQoXJ8r9z/uP1QT1\nG9JqpMB7o15NZnI4o/vGkZUSToLNeMHD14QZNdw0PIUx/eP5YlMxq3eUsGmfnSE9bIwfktiiA0pe\nDJE4ziM7NYJ1u0p54b81TIzvwdhxVwOgeDw4DxxEY7Ohiz//jGeCcD5Ot481O0uwmLUM7m6jX9co\npPGd6ZcZ3eJjg7UUs0HD9FGpDO8dw+JvjrH4m3y+21nKtKtS6Z5uCdq23iPz7Y4SVm0qptbpDaoQ\nToyPoKiobRLHuZiNGjoZw+iUEFy/KSsKVbWeQEIpqXCRHG8lMQpSYkwt3sM/wqRl6lWpXD0gPlDv\ns3GvvVXHbmuOcyaODz744KzrfD7fWdddTrqlRXB/Wjn/2O5maVh/qtYfZ/LwZFQ+mRN/e5XI8deh\nG39de4f5k1Ja6Z/7IybSgKIoHMyvQT5leA9Z9hcxJNqM+Hwy23IrkWUFOTC8h0JKjInUODOyolBW\n6cYaoWuVYp3z2XqwnB8Laykqc3LkeC2uepkhPWwM7m5Dq5EYmG1t85guRKLVyINTMtl5qJIl3+bz\n0tKD9MmI5OarUogM07F2VymffV9MVZ2H7mkRTLoy6YyTc0cmqVSB+cIbR+tti7ovyymDf376fVFg\n7LYRvWO4dlAiFnPoRYMt6ZyJw263n3PnkSNHtmgwHZG3vBy++JS7srP5LqsHq7eXMLSHjaQYE9rY\nmDOGWBdaT5HdyfJ1hWz7sYJhvWzceo1/RrYXlxw4Y9sx/eOYdlUqHp/Ca58cOmP9+MGJpMaZqar1\n8OSbu5AkFbYIHbFRBmIj9eR0s9IpIQxZVvD55DP2D5Xb46PY7qSwzElxmZNCuxOHy8e8n/s74n2/\nz87eI9Uk2owMyIpmeO8Y0uMvnRPqqVQqFX26RNE93cLXP5zg0++L+O3buwkzaqis9dAlKYy7JmSQ\nmRze3qFecqLCdfx8bLo/gTQMvfPdrjKuuiKWcQPjz9p9oLWcM3Hcd999bRVHh2Vf8iEoCrHTpjDD\nGs2YfnGBiqrahM6Y8w+2+DFlRUGWlXO26PgpKaty88mGQjbutaPXSowfnEj/TH9lp0ql4lfTsgKj\nzjaORhvR8Iek00o8PbsnEqCSGkemVWEKVCxLzBrXiZJKf/FDSYWbHwtqSIs30ykhjLzjddz/0idY\nLTpiI/1JJTbKQJ+MyKBWPx6vzPFyF4VlDortTiYOTUKjllj2XQHfbCsB/EPcJ0QbSbAZ8MkKaknF\n7Gs7o9eFNiTIpUKrkbh2UAKDe1hZvq6Q8pp6Zo3rRLe0iA7fMa+jO3XonU82FPHVD8dZs6OEUX3j\nmDg0sc3OGaKO4xwcu/fg2L6DqBsmorX6B3NsTBrbf6zg1dqujPTVMLWuDo35wnp+KopCWZUbWfHP\n7Fft8PDkG7tw1fsw6NSEmzSEGzWM6BPLkB423B4f3+0sJdykJa1GQ72jjnCTlnCTpkV/aRTFPxKs\nt2GCJ7NB3S5FOQDf7Sxly4FyxvaP59qGjmCnykyJOMue/iKGcw1TbtRrGNozeB6HxhFrAcKNGiaP\n6sLh/LJAUnF7ZJJsRqIj9Gz7sYL3/3uUqjpPYB9JUjG0ZwxxUQaG9LCRlRJBotWILVJ/Rl1FY8uo\ny1FkmI7bxp05T7dw8WIiDdx+XWeuy0ngk41FfL6pmOzUCLqlnf1voSWJxHEWcn099g/+gzY+DsvY\n0Wesz06LoFe8hm/oT9lHP3Ln1J4hD7K3I7eCvON1HDlex9ETdThcPgZ3t3L7dZ0JN2oY1suGSa+h\n1umhxuml1uGl8UKtoqaeJavzGz7pcOAzZzRMr3mi3MXbqw4TbtISZtQEWn+M6BNLeryZoyfqWLam\nwD/Ln+/kjH+zrulEl+Rwtv1YwesrD50x+dMvb+xK74xI6j0+NCEOmneh6pxevthSTNfkCHp2sjBu\nYDxXXRHbZhWCjfNjAMRGGbitR+dAWbaiKFQ7vIE7FpNBTXZqBNYIf51KotVIbNTJ9v9pcWbSQhyY\nUBCaK95q5K7xGdx2TXqzO3ReDJE4zqLqi6/wlpUR/+D9TfbXMOjU3D25O19tOMZHO6qY/6+93D2p\nC8kxJ5v4VdV5ONqQHBQFJl2ZBMBH6wo5Xu4iyWakX9do0uJNdEn0l/uqVKpzDuIXF2XgxV/2pdbh\nRW+2kHfsODVOD12S/Pv7FAW9Vo292k3e8Tp8Pn9P48bZ2hRFwe3xodVIGHRqNGoVWrWEvqH3bGyU\nnjH94homdvJP6qSWVHRP91/JfLapmA27yxiQbSUnO5qUWFOLFT+46n18vfUEX2w+jrveh0Yt0bOT\npWGu9hY5xEVTqVRBFZJZKRFkneOORxDaQmsN/3I2InE0wVNSStXnX2Ie0B9jdtZZt1MbDYwbnUlG\nZg2LPjlEYZmT5BgTK9YXsnZXKZUNPUdVKgIndvBfvVvM2gu6QlCpVIF+JomJViL17qD1iVYjD918\n9pjT48N4tGGE1KYk2UxMHnH2XsCdE8LIL3Hw9dYTfLnlOPHRBob2sDEuJ6HZ3+VUG/aUsfTbfGqc\nXq7oEsmkK5OabGcvCEL7O2vi2L17d0gf0DiV7OVCURTsHywBjYboqecfh8qxdx+2o8d45o6xgTFv\njHo1mSnhpMWZSY8zkxwbPPxzU0MOXCp6dY6kV+dIap1eth4sZ/P+co6eqAus37i3jOzUCCJDGF/I\n528/i1rtL05LjjFxw7BLq5mmIPwUnTVxvPLKK0Hvy8vLUalUhIeHU1NTg6IoWK1WXn755VYPsi05\ntu3AuXcf0TdPQWOxnHd718Efqfrya9LHjgb8yWFs//hWjrL9hTVU2I/oExuYy8Re7eatz/JQAZmp\n4eRkWembGRXohd9IVhR+OFDO8vWFjOnnr78Y3juGEX1Eb3xBuBScNXEsXLgw8PrDDz+ktraW6dOn\no9frcbvdfPDBB4SHX17tsWWXG/t/lqJLTiJi5PCQ9tGnpIAsU198HH1qSitH2DE1thSyRuj53e09\n2by/nE377fzzyyP8++uj/PKmrvRIt6AoCpv3HufN5XsoKHWSaDMSE+m/+xLNNAXh0hFSHcfKlSt5\n9dVXAzMA6vV6Zs6cyd13381NN7XfsOItrfLTz/BVVBJ75+2o1KFVNulS/BXe9QUFP9nEcar4aCMT\nhyYxYUgix0ocbNpnJ72hVdFXP5zgP9/mExOp587rOzMgK7rFh2oQBKH1hZQ4DAYDubm5ZGdnB5Yd\nOnQIvf7SLas/nePYMaq+/oawoYObNUGTxmZDpdeLHuSnUalUZzRF1Wkkfjm1Dz2SNe3WJ0QQhIsX\nUuKYPn068+fPp3///litVux2O1u3buXOO+9s7fjahKIoHPr7a0hGA9E33dCsfVWShC45CW9lZStF\nd/kYeUVsh5jbQhCEixNS4hgxYgSdO3dm48aNVFRUkJSUxJQpU0hOTm7t+NqEc+8+qvfsxTpzBuqw\n5rfoiX/wfiRt+ww2JgiC0NZC7seRnJzM1KlTWzOWdqONjyNt1q2ocgZc0P4iaQiC8FMSUuKora1l\n+fLlHD16FJfLFbSuo00deyG0ViuJvXpdcBGKr7qasvcWEz78Skzdu7VwdIIgCB1LSInjpZdewuv1\nMmTIEHS69p1ApCNSGY04du5CGx8vEocgCJe9kBLHwYMHef3119GKIpkmSVot2oR46gvyz7+xIAjC\nJS6kxJGamordbic+/sJ7RBcVFbFw4UJqa2sJCwvj/vvvJyHhzPGN1q9fz9KlSwPvn3zySSIjIy/4\nuG1Fn5yEc3/Lz80hCILQ0YSUOHr27Mn8+fO56qqrzjiJjx595pDjTXnttdcYN24cI0aMYM2aNSxa\ntIinn346aJtDhw6xZMkSnn76aSIjI3E4HIFOhx2dLiWZ2u8346uuQR1xefWoFwRBOFVIZ+X9+/dj\ntVrZtWvXGetCSRxVVVXk5eXx5JNPAjBs2DDefPNNqquriYg4OST1ypUrmThxYiA5mUyXzuio+tRU\n9Olp+OpqReIQBOGyFlLiOP3OoLnsdjvR0dFIkr+3sCRJREVFUVZWFpQ4CgoKiI2N5emnn8blcpGT\nk8PkyZMviXGMDF27kPjor9s7DEEQhFbX7HIg/7SaJ2eHa0wGLUGWZY4ePcpvfvMbvF4v8+fPx2az\nMXLkyJA/IzEx8YKPfzH7NlIUpU0TXUvE3NZEzK3vUosXRMxtpSViDilxlJeX88Ybb7Bv3z7q6uqC\n1n3wwQfn3d9qtVJeXo4sy0iShCzLVFRUYLMFz/Vss9kYPHgwWq0WrVbLgAEDyM3NbVbiuNC+GC0x\nFEb5so9x7tlL4qO/RtUGLdAuxeE7RMyt71KLF0TMbeVsMTc3mYR0u7Bo0SI0Gg1PPfUUBoOBP/7x\njwwYMIA5c+aEdBCLxUJ6ejpr164FYO3atXTq1CmomAr8dR87duxAURS8Xi+7d+8mLS2tWV+oPWkT\nE6gvLKLohT/jsZe3dziCIAitIqTEcfDgQe69917S09NRqVSkp6dz77338sknn4R8oDlz5rBq1Soe\nfPBBVq1aFUg6zz33HIcOHQJg6NChWCwWHn74YR555BGSk5NDbrXVEYQPyiH27rvwnCih6Lk/4tiz\nt71DEgRBaHEhFVVJkoS6YX4Ks9lMdXU1RqOR8vLQr6qTkpKYP3/+GcvnzZsXdJxZs2Yxa9askD+3\nozFf0QddYgIli96g9I23SX72t6jNl07rMEEQhPMJKXF06dKFbdu2kZOTQ58+fViwYAE6nY6MjIzW\nju+SpI2NJeGRX1FfWITabPI3KHC5kIzG9g5NEAThooWUOObOnRtoSTV79mxWrFiB0+lk/PjxrRrc\npUzS6TB0SgegZs13VH7+FbFz7ggsEwRBuFSFlDjM5lNmcdPpmDJlSqsFdDnSd+qESlJR/MKfsU6d\nTPjI4ZdE3xRBEISmiPk724A+NYXEeY9i7JaN/YMllL75D2SXu73DEgRBuCAicbQRtdlE3L2/IGrS\nBOq2bsOdl9feIQmCIFyQS2MEwcuESpKIvG4c5pwBaK1WADwnStDGxbZzZIIgCKEL6Y7jyJEjrRzG\nT0tj0nAfOULBM3/Avvg/KF5vO0clCIIQmpDuOJ599lmio6MZPnw4w4cPJyoqqrXj+knQpaQQcdUI\nqv+7GvfRY8TedTsa8bMVBKGDC3nIkWnTppGbm8sDDzzA73//e9asWYPbLSp4L4ZKrcZ68xRi7rqd\n+sIiCuf/H859+9s7LEEQhHMK6Y5DrVYzcOBABg4ciMPhYMOGDSxfvpzXX3+dnJwcxo4dS3Z2dmvH\netkK698PXVIiJYvewH3sGMZu4mcpCELH1azKcZfLxaZNm1i/fj12u52hQ4dis9n461//St++fbnr\nrrtaK87Lni4+nsTH/gdVw4yHNes3ILvcmHr3RHvaKMKCIAjtKaTEsXXrVtasWcO2bdvIzs5m9OjR\nPProo+h0OgCuvfZa7r33XpE4LpLU8PMEcOzcjWPHTsqXLEWblIipVy/MV/RGn5bajhEKgiCEmDje\nffddRo4cyaxZs5qsGA8LC2P27NktHdtPWtw9c/CUlOLYtQvHjl1UffElnpITxM25EwDn/gP4Glpn\nCYIgtKWQEscLL7xw3m3GjBlz0cEIwbSxMVjGjMYyZjS+ujpkpxMAT2kZx196mZJXFmHoloWpd29M\nPXuIuc4FQWgTIbWqev7559m3b1/Qsn379oWUUISWoTabA3UdmqhI4h/4JXFXj8F9rICyf77Lscee\nwLFrNwCKzxc0va8gCEJLCumOY+/evTz88MNByzIzM/nTn/7UKkEJ56bSaDB2yyZxzGj046+jvqAQ\nx85d6DulA1D97XdUr16DqXdPjN27YeiULoZ0FwShxYSUOLRaLS6XC5Pp5IRELpcrMLmT0H5UKhX6\nlGT0KcmBZdoYG9rYGH8C+fobUKnQJSf5W21JEr66OiSTSYzQKwjCBQkpcfTp04dFixbxi1/8ApPJ\nhMPh4I033uCKK65o7fiEC2Dq1RNTr57ILjfuvDxceUeQa2pRSf6SyZJFb1BfWIi+Uzr6Tp0wdO6E\nPj0NyWBo38AFQbgkhJQ4brvtNv76179yxx13EBYWRm1tLVdccQVz585t7fiEiyAZ9Bi7ZZ/RoTD8\nyqE4Dx7EfTgP527/vOjG7t2In3sfAHXbd6BLTEATEyPuSgRBOENIiSMsLIx58+ZRUVGB3W7HZrMR\nGRnZ2rEJrSQsZwBhOQMA8DkcuI8cDXQ8lJ1OSha9AYqCFBaGvlM6hs6dMPXuhS4xoR2jFgSho2hW\nz/GoqCgiIyNRFAVZlgGQJDGlx6VMbTJh6t4t8F6l15P0xGO48vJwHz6C+/BhKnbtRjIY0CUm4Kuu\npvaHbRizs9DGx4k7EkH4CQopcZSXl/PGG2+wb98+6urqgtZ98MEHrRKY0D5UkoQuKRFdUiIMuxIA\nX20dKrX/AsH5Yy7li/8DgDoyEmN2JsZu2Zh69xJ1JILwExHy6LgajYannnoKg8HAH//4RwYMGMCc\nOXNaOz6hA1CHmQPNecP69yP52aex/nwGhoxOOHbtofStd5Ad/s6JrkOHcezegyxGThaEy1ZIdxwH\nDx7kb3/7GwaDAZVKRXp6Ovfeey+/+c1vGDt2bGvHKHQwWpsN7TAbEcOuRJFlPEXFaKL9Q9FU/3c1\ndVu3gVqNoXMnDNlZmLpno09Pb9+gBUFoMSElDkmSAn02zGYz1dXVGI1GysvLWzU4oeNTSRK65KTA\ne9usWwgfNhTn/gM49+2ncsVKHNt3kPT4owAUf/Y5NS4nWqsVjTUatcUSaCYsCMKlIaTE0aVLF7Zt\n20ZOTg59+vRhwYIFmknwBAAAIABJREFU6HQ6MjIyWjs+4RIj6XQnmwDfdAO+mhp8VdWB9Uf/+S98\ndY6TO2g0RAwfhnXaFACqvvov6kgLGqsVrdWKFB4mKuAFoYMJKXHMnTs3MPbR7NmzWbFiBU6nk/Hj\nx7dqcMKlTx0ejjr85OCLOW+/wbHde/Da7Xjt5Xjt9sAdi+x0Ur50WdD+Kp2OyPHXEXnNWOT6emrW\nrEVjs6JLSEATYxN3K4LQDs6bOGRZ5q233uLuu+8GQKfTMWXKlFYPTLg8SToduvg4dPFxZ64zGklb\n8KdAQvGU2f2JpaH/iNduD0osKr0OXVISUROux9gtG8XrRZHloHlNBEFoeedNHJIksXPnTlFcILQJ\nyWA42Rz4NNr4eFKf/yPesjLqC4uoLyjAnV8ADfVvzv0HOPHKIrRxcehSktCnJKNLTkbfuZNIJoLQ\ngkIqqho/fjyLFy9m2rRpaDTN6jMoCC1GpVKhNptQm1ObnAlRY7MSOe5q3PkFuA7mUrdpCwBJv5mH\nLikR5779uHIPoUtJRpechMZqFRdEgnABQsoCq1atorKykpUrVxIRERG07pVXXgnpQEVFRSxcuJDa\n2lrCwsK4//77SUhoegiLoqIiHnnkEa655hpuu+22kD5fEHTx8egmTQi891XXUF9YiLahWMyVd4TK\nzz6Hhvo6yWhEn9GZ2Dl3iDsSQWiGkCvHL9Zrr73GuHHjGDFiBGvWrGHRokU8/fTTZ2wnyzKLFi1i\n4MCBF31M4adNHRGOMeLkAI9R11+LZezoQDFX/bF8PPbyQNIo/de/8VVVY8zsiiGzK7qUZFH5Lgj/\nv707D4/xXB84/p2ZzGTfE8Qe1FohVCMobTni9Kj2OPZyDqdBbV0ouvBDLaG0liyWkqqWluiGNlpL\nqaCcqtiKSkSlUmSSyD5JZvn9MfI2I5NkhmSSyPO5rl417zb3Ox5zz7vdtxkWJY727ds/0JtkZmaS\nlJTE3LlzAejVqxfR0dFkZWWVOoL56quv6NKlCxqNBo1G80DvKwj3kqtUOPg3x8G/eal5ChcXChIS\nST9/AQCZgwOuPbrT8BXjDyeDwSBObQkCFiaO8upRDR8+vML109LS8PLykgoiyuVyPD09UavVJonj\n2rVrnDlzhnnz5rFz505LQhOESuP1/CC8nh+E9k4mmitX0PyWgMLdHTC2402eMx9Vk8Y4tm6FQ+vW\nqBo3EkckQp1kUeJIS0szeX3nzh1+/fVXHn/88UoLRKvVsmHDBiZPnvxAFXcbNix9N44t1q0uIuYq\n0LAhlKgYDFDPw4P8oG5knrtA+udfAaBwdqLFhPHUe7I3Oo2Gwow7ONTzRVYDOmPW+M/YDBGzbVRG\nzBYljsmTJ5eaFh8fT1xcnEVv4u3tTXp6Onq9Hrlcjl6vJyMjAx8fH2mZO3fucOvWLcLCwgDIzc3F\nYDCQn58vPUNiiZSUFIuXLalhw4b3vW51ETHbRsOGDUnNzsbp+UE4PT8I7Z07aH5LQHPlCjlKO7Qp\nKeRd+JVbEWtBoUDp441dvXoo6/ni1qc3Sl8fDDodyGQ2OUKprZ+xiLnqlRWztcnkvu+tDQgIYOXK\nlRYt6+7uTvPmzYmLi6N3797ExcXh7+9vcprKx8eHTZs2Sa937NiBRqMRd1UJNY6dh4dJMywAVaOG\n+IwZRdHtVIpu3aYoNRXNpcu4BBmPyrOP/0T6js+xq+eL0tcXZT1flPXr4dwlUJSjF2odixLHrVu3\nTF4XFBQQFxdncsRQkfHjxxMZGcnnn3+Os7MzU6dOBSAsLIxhw4aJuldCrWbn4YFrj2CTaYa7zc4A\nVA0b4tq7F9rbqRT9+Sd5586DTodTpwAA7ny/n9yff0FZv97d/+qjbFAPVSNxHUWoeSxKHC+//LLJ\na5VKhb+/P1OmTLH4jRo1asSSJUtKTX/zzTfNLj9s2DCLty0INVHJL3yHFv44tPCXXht0OrTpGSic\nnYHiml4uFCQlkXvqFzAYkKlUNFu5HIDMAwcpSlUbn4pvUB9l/XooPDxEUhGqxQPfVSUIgvVkCgVK\n37+O2F2Dg3ANDgJAX1iI9nYq2qwsKTEU/nmT3FOnMZS4RV3ZqCGN5xh/eGX/dAKZXI6yfn20Hp42\n3BOhLrIocVy7dg0XFxeTU1NqtZqcnByaiwY9glCp5CoVqsaNUPFXnxPf0aPweWEkuqws4zWUW7eQ\nyf462rizJxbt3bsfUwC5szPOj3XBZ4TxyD3nxP+QOzthd7cPinhSXngQFiWO8PBwZs2aZTJNq9US\nERHBihUrqiQwQRBMyWQy7NzdsXN3x7H1IybzGs97myK1mqKbt3DUFJCelISyXj3AeK0l9eOtoNNJ\nyyvc3HB7sjcefw/BYDCQffTY3eZa3th5eSITNemEclg0OtRqNfXrm5bBbtCgAampqVUSVGUzGAxo\nNBr0en2ZT/7euHGD/Px8G0f2YETMtlFrYnZ3R+bujszREefOATgU360lk9Fk0YK7PVD+Kldv5+UF\nGGt6pW397K/tyGQoPNzxHPgMrj2C0efnk/tLPHbeXsb/PEViqess+tv38vLi6tWrtGjRQpp29epV\nPD1rx7lUjUaDUqkst7KvUqmsdeUkRMy2UdtiViqV6HQ6NBoNjo6OxiMVD3fsPNyhZYtSyytcXWiy\neIGUULRp6WjVadJT80W3bqP+ZNtfK8hkKNzd8Rk5DKeAjmgzMsi78Ct2Xl4ovb1ReHkiVypttbtC\nNbC4rPry5csZNGgQ9evX59atW+zevZvBgwdXdXyVQq/Xi3LwQp1iZ2dHQUGBRcvK5HLsvLzuHoE8\nUmq+qkljGi+aLzXY0qalo01PlxJLQdLvpkcsgMLdjXoTx+Pg35zCGyloEhKM7+HpaUwsdxOaUDtZ\n9G3ar18/nJ2dOXjwIGlpaXh7e/Pvf/+b7t27V3V8lUIMUKEuqqxxL1MoUN7tAW8usTh16kiTxQvQ\npqVTlJ4uJRg7d+MDvvmXfyM95nPTbTo40Oit2Sh9fci//Bt/HPuJPDvF3cTihZ2He40o3SKYZ/HP\n8ODgYIKDgyteULBIdnY2Q4YMYeDAgZVStr6mW7ZsGRcvXqRly5ZSleRiO3fupG/fvvd16lOtVrN4\n8WKLqhiEhoYSGRmJvb291e9jib1793L8+HEWLFhQ7nJxcXF4e3vTrl27cperLWQKhXTEYu4ZeLcn\ne+PcNRBtegba9HR06RnGZ1juJhbN5d+4Gfud6UpyOc3efxe5vT05J3+m8MYNFK6uyOztkTvYI3d0\nxOnRDgBoMzNBp5fmiYRT9SxKHNHR0fTs2ZM2bdpI0y5fvszx48cZO3ZsVcX2UNu/fz/t27fn4MGD\nvPTSSyir+JywTqdDUU3/oNLT0/nxxx/ZvXs39vb2FBUVmczfuXMnXbt2NZs4im9oKOvXs4+Pj8Wl\nbzZu3Gh98FUgLi6ONm3aPDSJoyIyuVy6Gwz/5qXmew4aSNtx/yH5wgW06RnGxJKVhfxugi9IukZW\n3FHQaqV15M7ONFuxFIC0z2LIiz/z1wbt7FDWryc945K243MKU1KQ29sbk4uTI8p69XB/+knj9pOT\njddtXFxQODsjE9dnKmRR4jh69GipmlEtWrRg+fLlInHcp9jYWCZOnMi2bds4evQoTz75JACpqamE\nh4dz48YNAJ5++mleeOEFcnJyiIyM5PLly8hkMgICAnj99ddZunQpbdq04Z///CeAyeulS5eiUChI\nTk4mLy+PjRs3smjRIpKTkykqKqJRo0bMmjULV1dXAL799ls+/9x4SkGpVLJkyRK2bNlCgwYNGDFi\nBABXrlxh4cKFfPTRR6W+zL/77ju2b9+OTCajYcOGTJ8+HXt7e6ZPn45Go2HChAk888wzJtfGPvnk\nE9LS0pg3bx4qlYo5c+Zw6NAhrl27Rm5uLrdu3SIyMpJPPvmEM2fOUFRUhLu7O7NmzaJBgwbcvHmT\niRMn8vXXXwPw1FNP8eKLLxIXF0dWVhYTJ06kT58+0rxvv/0WR0dHRowYQf/+/Tl16hRpaWkMHz5c\n+gzPnj3LqlWrkMlkdO7cmaNHjxIWFoa/v7/J/hYVFbFmzRpOnz6Nu7s7rVq1kuZdvXqVVatWodFo\nKCwsZODAgQwZMoSTJ09y7NgxTp06xTfffMPQoUPp1q0bCxcuJDc3l8LCQrp3785LL7304IOsFlHY\n2xs7ODZoUGqe9/AheA0djKGgAH1BAXpNgUkScXuqD06PdkBfUCAtIyv5nIpCjkGrRZubi15TgD4/\nzyRxqLdso/CPP6TFZQ4OOLVvR73x/wUg/atdGLRaY2JxcUHu4ozS19dYRbmOsihxyGQy9CXq7oDx\nl6DhbgvO2iT7pxPkHPup1HSZTPbA++PSozuu3YMqXC4xMZGsrCy6dOlCeno6sbGxUuJYsmQJQUFB\nvPPOO4CxCRZAZGQkjo6ObNy4EblcLk2vSEJCAqtWrcLR0REwdnN0v3tRc9OmTXz66adMmDCB+Ph4\ntm7dSnh4OF5eXuTn56NQKPjnP//JW2+9xfDhw5HJZHz55ZcMGjSoVNJISkrigw8+YP369Xh7exMd\nHc2aNWuYN28eS5cuZeLEiWzcuBGlUmlyxDF69Gj27NnDggULTL6YL168yIYNG6RYR40axaRJkwD4\n5ptv2LBhA//3f/9ndp+dnZ1Zt24d586d45133pESx70KCgqIjIzk5s2bjBs3jgEDBqBQKFi4cCFz\n584lICCAI0eO8MUXX5hdf/fu3dy8eZPNmzej1Wp55ZVXaHD3i69BgwasWLEClUpFfn4+kyZNolu3\nbjz++OP06NHDJNkXFhayZMkSHB0d0Wq1zJo1i5MnT1Zq24LaTiaXI3N0RH53HJfk2PoRaF362ksx\n73/9s9S0kv/WvUcNR3fnDrqcXPQ5OehycqRblQHyL16i6OYtDIWF0jTnrl2gm7HIZermj7Hz8ca+\naRNUTZuicHd76K+rWpQ42rZty2effcbo0aOlsugxMTG0bdu24pWFUr799lv69++PTCajd+/ehIeH\nk5qaiouLCxcuXDB5qLL4i/P48eOsX79e6lVSPL0iffr0kZIGGI8K9u/fj1arRaPR0LhxYwB++ukn\n+vfvj9fdfzDF6zRr1oyGDRty8uRJ2rdvz7Fjx8yW2T99+jRBQUF4e3sD8OyzzxIaGmrtRyMJCgoy\n2ccTJ07w1VdfkZ+fj67Eg2zmPP3004Cxc6VaraawsBCVmSeli5dr0KABrq6upKamUlRUhL29PQEB\nxuKDTzzxBC4uLmbfJz4+nv79+2NnZ4ednR1/+9vfOHfuHGC8BXzt2rUkJiYik8lQq9UkJibSrFmz\nUtvR6XSsW7eO8+fPA8ZTewkJCSJxVKGSX+zmukGW1OhN48PP+sJC9Dm56HJzpNNZeo2Ggt+vk3Py\nf1Ive4WbKx7/eAa33r0w6HToMjNReHo+VMnEosQxbtw46Vejj48ParUaT0/PUk+T1wau3YPMHhXc\n+0u4qhQVFXHgwAGUSiXff/89YHwK/7vvvuNf//qX1dtTKBQmR4OFJX4VASZJ4+zZs+zatYuIiAg8\nPDzYv38/e/bsqfA9Bg8ezNdff83vv/9e7hdpZSoZ982bN4mMjGTdunX4+flx/vx5Fi1aVOa6xUmi\n+JpOWYmmZDKRy+UVJiRrbNy4ES8vL9544w0UCgUzZ84s9XdTLCYmhuzsbNauXYtKpWLFihVlLitU\nH7lKhdxLhZ3XX9fi5A4ONJ73NnpNAYU3blBwPZnC69ex8/QAoDDlT1KWLEPu4oJ9k8aomjbBvmkT\nHFq3RuHiXF278sAsShze3t4sW7aMhIQE6XbckudzBcsdPXqUJk2aEB4eLk27cOECYWFhjB49mg4d\nOhATEyNdU8jMzMTd3Z3g4GC2b9/OtGnTkMlkZGZm4uPjQ6NGjbh8+TJg7NQYHx9Px44dzb53Tk4O\nzs7OuLm5UVhYSGxsrDSve/fuLF++nGeffdbkVJVKpSIoKIioqCiuXLnC0qVLzW47MDCQbdu2kZ6e\njpeXF3v27KFr164WfSbOzs7k5OSUOT8vLw+lUomXlxd6vZ5du3ZZtN370aRJEwoKCjh37hwdO3Yk\nLi6uzNgCAwPZt28fTz/9NFqtlgMHDlDvbpmPnJwcWrRogUKhICkpibNnz9K3b18AnJycyM3NlbaT\nk5ODt7c3KpWK1NRUjh07xqBBg6psH4XKJ3ewx6FlCxzuecDSzt0N7xFDKbj+B4XXk8ncdwD0eupP\nm4xT+3YUXLvGnb37jCX49Xrp/77/fgE7Ly9yTv5M5sEfTOah0+M34xUUbm5k7j9I5r4DOAV0xOeF\nETbbX4tvx5XL5bRu3RqA69evs3XrVuLi4li/fn2VBfcwio2NpV+/fibTOnTogMFgID4+nrfeeovV\nq1czbtw45HI5/fr1Y+TIkUyZMoWIiAjGjRuHQqGgU6dOzJgxg3/84x/Mnz+fsWPH0rhx43Lv1Hn8\n8cfZt28fY8aMwd3dnYCAAC5dugRA586dGTVqFK+//joymQyVSsXixYulXvEhISGcOHGizL4p/v7+\njB8/Xlrfz8+P6dOnW/SZDB48mHfffRd7e3vmzJlTan6LFi3o06cPY8eOxd3dnaCgIM6ePWvRtq2l\nUql4++23WblyJTKZjE6dOuHp6Ymzc+lfhwMHDiQxMZH//Oc/uLu706ZNGzIyMgAYM2YMS5YsITY2\nlsaNG0unvgD69+/P0qVLOXToEEOHDmXw4MEsWLCAcePG4evrS2BgYJXsm2B7Cjc33Pr0ll4biooo\nTPkT5d0SToV/3kSrVoNcDnK5sRqyXC5dg5GplChcXKTpKOTG4pZy49G0sn49nAIeRdWsiU33S2aw\n8IpwVlYWcXFxHD58mGvXrtG2bVsGDBhQ457tMNcWMS8vDycnp3LXs9Wpqspky5hff/11Bg4cKF3E\nv1+14XMuOV5Onz7NsmXL2LZtm3R9qaYr/owtGfc1xcPUhrUms0nrWK1Wy88//8yhQ4c4c+YMDRo0\noGfPnqSmpjJ9+nSLL9AKtdfly5d55513aNWqFb179654hYfAjz/+SExMDAaDAZVKxbx582pN0hAE\nWyg3cYwfPx65XE6fPn0YNmyYVOSw+KKu8PBr06YNW7dure4wbGrAgAEMGDBAel0bjpIEwZbK/RnV\nrFkzcnNzSUhIIDExsdwLmIIgCELdUO4Rx/z580lNTeXw4cPs3r2bDz/8kICAAAoKCir11kVBEASh\n9qjwripfX1+GDBnCkCFDuHTpEocPH0YmkzFz5kyeeuopRo8ebYs4BUEQhBrCqiYVbdu2pW3btowb\nN46TJ0/y448/VlVcgiAIQg11X92NVCoVvXr1olevXpUdjyAIglDDiXsMq0l2djYhISEmT5A/zJYt\nW8bYsWPLLExoraVLl/Lll18CsGvXLmJiYswut3fvXubNm1fh9uLi4rh48aL0+vLly+WWNakMI0aM\nICkpqdxlcnJy+PTTT6s0DkGwlkgc1aRkPw5b3OpZnTczFPfjiI6Olqr+VqZBgwYxdOjQB9pGXFyc\n9BQ9GG9DNvcUu63l5OTw2WefVbygINiQaMRdTUQ/DqN9+/bx448/snDhQsCY4IYPH054eDj5+flm\ne1rca/PmzVLp8srqkVG/fn3Wrl0rldQxt2+enp7s3buXAwcO4OLiQlJSEi4uLrzzzjtSleGSivt8\nAHTq1Mlk3tq1a832G1m1ahU5OTmEhobi4OBAREQEO3bs4ODBg+h0OlQqFa+99pqoHSfYVJ1LHMcv\nqDl2Xl1qemX04+jxqA/BHXwqXE704/jrCOuJJ54gIiJCKuZ44sQJmjZtip+fH3l5eWZ7WpgrTV6s\nsnpkxMfHS9u8evVqmfsGcOnSJTZt2kS9evVYsWIFX3zxRamS8oWFhSxcuJC3336bzp0788MPP/DV\nV19J88vqN/Lqq69Kn12x/v37M2zYMABOnTrF+++/T1RUlCXDQRAqRZ1LHDWB6MfxFwcHB3r16sX+\n/fv517/+xd69ewkJCQGs62lRrLJ6ZJT0yy+/lLtvjz76qFQVt127dpw6darUNpKTk7G3t6dz586A\nsRvhe++9J823pt/Ib7/9xtatW8nKykIul5OcnFzu8oJQ2epc4gjuYP6oQPTjKFtV9+MYMGAA4eHh\n9OvXjzNnzvDWW28B1vW0sERlb69Yyb4eCoXC4utJxUdt1vQbKSoqYt68eaxevZrWrVujVqsf+PqO\nIFhLXBy3seJ+HDExMXz22Wd89tlnLF++nL179+Lo6Cj14yhWfEqquB9H8em04unm+nGUpaJ+HN9/\n/z3p6ekA5OfnS1+qQUFBJCcnExMTw/PPP29224GBgZw4cUJa35p+HB07diQvL48PPviAXr164eDg\nIMXr6+tr0tOiIsU9MnQ6HQUFBRw4cMBk/8va3r09Mkrq0qXLfe9bseI+H8XvefjwYamET3n9Rpyc\nnEwqNRQWFqLT6aQjnOJe64JgS3XuiKO6iX4c5oWEhEjXDoqV19OiLJXVI6P+3X4JYOwH8iD7Bsaj\nkrlz57Jq1Srp5obi9yiv34ibmxv9+vXjv//9L66urtIYeOmll3Bzcyuzn7ogVCWL+3E8qJSUFCIj\nI8nJycHFxYWpU6fi5+dnsszOnTs5duwYcrkchULByJEjpXPC1rzPvUQ/jgdXl/px3Ku2xSz6cdjG\nwxSztf04bHaq6oMPPiAkJITVq1cTEhLChg0bSi3TqlUrwsLCWLFiBZMmTWLVqlWi93I1u3z5Mi+8\n8ALOzs51ph+HIAjls0niyMzMJCkpSSpR0qtXL5KSksjKyjJZrnPnztjb2wPGu3kMBgPZ2dm2CFEo\nQ3E/jgULFohmRoIgADa6xpGWliadKwdj/3JPT0/UajVubm5m1zl8+DANGjSQboG0lLlDrhs3bqBU\nKitc15JlahoRs23UtpiVSiWOjo5Wn4KoTrUp1mJ1NeYaeXH8119/Zfv27cydO9fqdc2dv8vPzy/1\nwNq9att5bBAx20pti7k43vz8/FpzDv5hul5Qk9Wqaxze3t6kp6dLzxvo9XoyMjLw8Sn9PMVvv/1G\neHg4M2fOrJXZXBAE4WFnk8Th7u5O8+bNiYuLA4wF5fz9/UudpkpISGDlypVMnz5d6m8uCIIg1Cw2\nu9o5fvx49u7dyyuvvMLevXsZP348AGFhYSQmJgLG2kmFhYVs2LCBmTNnMnPmTK5fv26rEG3GknLa\ntpCZmcnUqVMJDQ0tVYH1Qct5Hz16lHXr1lW4nC3Kl5cswV6enTt3Ss98CIJQNptd42jUqBFLliwp\nNf3NN9+U/hwWFmarcASMBfJcXFyIiIgoNa+4nPfIkSPNrqvT6VAoFGVuu2fPnvTs2bPCGGpK+XIw\nJo6uXbvi6elZ3aEIQo1WIy+O11Vlle4+f/48a9asQa/Xo9VqGTNmDH379uXrr79m+/btKJVKDAYD\n8+bNo2nTpibb1Ol0bNiwgZMnTwLGp8cnTJjA2bNnWb9+Pbm5uYSGhvLyyy+bPEltrpz3q6++SqtW\nrfj1119xc3Nj8eLFvPHGG2RlZVFQUEC7du2YPn06SqWSvXv3cvz4cRYsWEB8fDwRERG0a9eOixcv\nYjAY+L//+z+aNWtGfHy8VL785s2bTJw4kWeffZYTJ05QUFDAzJkz6dixIwBffvkln3/+OS4uLgQF\nBfHVV1+ZLbmRmppKWFgY6enpNGjQwOTGiP379/P555+j1WoBeOmll+jatSuffPIJaWlpzJs3D5VK\nxZw5c0hLSyM6OpqioiK0Wi2jR4/m6aefrvS/d0Gobepk4vjz/dWlprk9/hjOvXqiLyzkVsTaUvNd\ngoNwDe6OLieH2xs2lZrv2rsXLo9ZV7+opPLKkn/66acMHz6cvn37YjAYpJpKUVFRfPTRR3h7e1NY\nWGhS7LDYnj17SEhIkB64nD17Nnv27OG5555j3Lhx0pf7vcyV8wb4888/CQ8PR6FQYDAYmDNnDu7u\n7hgMBsLCwoiNjWXQoEGltnft2jVmz57NG2+8wYcffsjHH39s9kgjKyuLDh06EBoayr59+1i/fj0R\nEREkJiaybds2PvjgAzw8PMrtnBgeHk6nTp34z3/+Q0pKCqGhoTz++OMAdOvWjb59+yKTybh+/Toz\nZswgJiaG0aNHs2fPHhYsWIC/vz9gvKljzZo1ODg4cOvWLSZOnEi3bt2k/iWCUFfVycRRE5VXljww\nMJCPP/6YlJQUunbtSvv27QFj8b2lS5cSHBxM9+7dzd6FdurUKQYMGCA9h/D3v/+dI0eO8Nxzz91X\nnH379pVOUen1enbs2MGJEyfQ6/VkZ2dLBQrv1aRJEx555BHAWHr82LFjZpdzdHQkODgYgPbt27N2\nrTGJx8fHExQUhIeHh7Qf+/fvN7uN+Ph4pk2bBhhvM+zSpYs0LyUlhUWLFpGamoqdnR3p6emkp6eb\nbbx0584d3n33XW7cuIFcLic7O5vk5GTp8xeEuqpOJg6/6a+UmlZ877tcpTI7v5jCxaXc+VVhyJAh\nBAcHc+rUKcLDw3nsscd48cUXWbJkCefOneP06dNMnz6d1157jaCgoCqNpWSZ9gMHDnDu3DnWrFmD\nk5MTn3zyCX/88YfZ9SwtPV7yQTtrSpRbatGiRUyaNIlevXqh1+sZMGBAmWVtVq5cSY8ePQgLC5NO\nEYoSOIIgyqrXGOWVJU9OTqZRo0YMGjSIwYMHc+nSJXQ6HSkpKbRr145Ro0bx2GOPceXKlVLb7dq1\nK9999x1arVbq+/HYY49VGM+95bzNycnJwd3dHScnJ3JyckxKmFe2Tp06ceLECamc/HfffVfmsoGB\ngezduxcwnlr75ZdfTGIu7ggYGxtr8mCfs7OzVOq85LIymYyff/5ZaucrCHVdnTziqAlmzJhhcldS\ndHR0maW7v/jiC06fPo1SqUSpVPLyyy+j0+lYvHgx2dnZyGQy6tWrJ93iXNLAgQO5ceOGNK9bt278\n4x//qDA+c+XTOD60AAAQG0lEQVS879W/f3+OHj3Kv//9bzw9PQkICKCgoOB+P5JytWrVihEjRjBl\nyhScnJzo0qVLmQ2lpk6dSlhYGAcOHKBBgwYmFZanTJnC3LlzcXV1pVu3bibPEg0ePJh3330Xe3t7\n5syZw4QJE1i1ahUfffQRbdq0Ec8WCcJdNiurbiuirHrN9iAxl/x73Lx5Mzdu3ODtt9+uzPDMqm2f\nsyirbhsPU8zWVukQRxxCrbFhwwbOnz+PVqvFz8+PGTNmVHdIglAnicQh1BqvvvpqdYcgCALi4rgg\nCIJgJZE4BEEQBKuIxCEIgiBYRSQOQRAEwSoicQiCIAhWEYmjGtSGfhz3q+S+LV++nLNnz5pd7n57\nZOzatYuYmJhKidWcmzdvWlTHKyEhgR9++KHK4hCEmkzcjluHldePozLMnDnzgbdxb48Mc5V3q0NC\nQgLHjx/nqaeequ5QBMHmROKoQWpSP47ly5fj7+/PkCFDAGPZ97fffputW7dy4MABsz0t7vXqq68y\nfPhwgoODpR4ZGRkZ1K9f/757ZBw6dIj8/HwmTZpU5r4pFAqWLl2KSqXijz/+4Pbt27Rv354333zT\n5H2Lffnll+zcuRNnZ2e6d+9u8tm98cYbZGdno9FopH4jeXl5fPjhh9JnFxAQwMsvv8yiRYtITk6m\nqKiIRo0aMWvWLFGCXXgo1cnE8d72S6WmPd7ehyc6+lBYpCP8i9LFAoM7+NDjUR9y8opYvzux1Pze\nnXzp1tb7vmOqaf04BgwYQHh4uJQ4YmNjCQkJQSaTldnTojzFPTJCQ0P5/fff77tHhqX7VvyZvvfe\ne8hkMsaPH8+pU6dKFXhMTExk69atbNiwAS8vL1auXCnNk8vlzJkzBx8fHwoLC036jZj77KZNm4a7\nuztgbIP86aefMmHChHI/F0Gojepk4qiJalo/jo4dO5KXl8fVq1dp1qwZBw8elE5pWdPTolhl9ciw\nZt969eollXNv3bq12Ro98fHxdO/eXXqvZ599lkOHDgF/9Rs5efIkOp2u3H4jYDxi3L9/P1qtFo1G\nQ+PGjcuNXxBqqzqZOGYMb1tqWnFhOJVSYXZ+MRcnZbnzq0J19eMICQlh7969dO7cmaZNm0rlyK3p\naWGJyt5esZI9QORyudW9PYr7jURFRaFUKsvtN3L27Fl27dpFREQEHh4e7N+/nz179jxQ/IJQU4m7\nqmqImtaPA4xl0w8ePMg333zD3//+d2l6eT0tytu/yuiRUVn7Vqxz586cOHFCunPr22+/NYmrrH4j\nTk5O0inD4mWdnZ1xc3OjsLCQ2NhYq+IQhNqkTh5x1AQ1vR8HQP369WnWrBlnzpxh7ty50vTyelqU\npbhHxujRo6lfv/5998iorH0r1rJlS0aNGsW0adNwcnIyuThe3G9k1KhReHh4mPQb6dKlCzt27ODF\nF1+kU6dOTJ48mX379jFmzBjc3d0JCAjg0qXS19IE4WEg+nHcVdt6LoCI2VZqW8yiH4dtPEwxW9uP\nQ5yqEgRBEKwiEocgCIJgFZE4BEEQBKvUicTxkF3GEQSLiHEvVJU6kTjkcrlUzkIQ6gKtVotcXif+\neQvVoE7cjuvg4IBGo6GgoMBsrSIAR0dH8vPzbRzZgxEx20Zti9nR0ZGioqJyn3IXhAdRJxKHTCbD\n0dGx3GUeplvrajIRc9WrbfEKtY/NEkdKSgqRkZHk5OTg4uLC1KlT8fPzM1lGr9cTHR3NmTNnAHj+\n+efp27evrUIUBEEQLGCzk6AffPABISEhrF69mpCQEKmiaUlHjhzh1q1brF69msWLFxMTE8Pt27dt\nFaIgCIJgAZskjszMTJKSkujVqxdgrFqalJREVlaWyXLHjh2jb9++yOVy3Nzc6NatGz/99JMtQhQE\nQRAsZJNTVWlpaXh5eUl3ecjlcjw9PVGr1SZ1idRqNT4+PtJrHx8f1Gq1Ve9l7aPzlbVudREx20Zt\ni7m2xQsiZlupjJjF/XqCIAiCVWySOLy9vUlPT5c61On1ejIyMkyOLqD0Eca9RyCCIAhC9bNJ4nB3\nd6d58+bExcUBEBcXh7+/f6ly3MHBwRw4cAC9Xk9WVhb/+9//TMpcC4IgCNXPZmXVb9y4QWRkJLm5\nuTg7OzN16lQaNmxIWFgYw4YNo2XLluj1ejZt2sTZs2cBeO655+jXr58twhMEQRAs9ND14xAEQRCq\nlrg4LgiCIFhFJA5BEATBKiJxCIIgCFYRiUMQBEGwSp2ojlusthVazM7OJiIigps3b2JnZ4efnx8T\nJkwodRtzZGQk586dw9XVFTDe1jx48ODqCBmAKVOmoFQqUSqVALzwwgt07tzZZJmCggKioqK4evUq\nCoWCMWPG0LVr1+oIl9u3b7N8+XLpdV5eHnl5eXz44Ycmy+3YsYPvv/8eT09PANq0aUNoaKhNYtyy\nZQsnTpwgNTWVFStW0LRpU8CyMQ3VM67NxWzpmIbqGddlfc6WjGmonnFtLmZLxzTc57g21CHz5883\nHD582GAwGAyHDx82zJ8/v9Qyhw4dMixatMig0+kMmZmZhokTJxpu3bpl61ANBoPBkJ2dbTh//rz0\nesuWLYaoqKhSy0VERBhiY2NtGVq5Jk+ebPj999/LXSYmJsawdu1ag8FgMKSkpBhCQ0MN+fn5tgiv\nQh9++KFh48aNpaZv377d8NFHH1VDRAbDxYsXDampqaU+W0vGtMFQPePaXMyWjmmDoXrGdVmfsyVj\n2mConnFdVswllTWmDYb7G9d15lRVbSy06OLiQocOHaTXjzzyiNW1u2qqY8eO8be//Q0APz8/WrZs\nyenTp6s5KmPnvCNHjvDUU09Vdygm2rZtW6qKgqVjGqpnXJuLuaaPaXMxW6M6xnVFMVfFmK4zp6ps\nWWixKuj1evbt21fmYe+ePXvYt28f9evXZ9SoUTRu3NjGEZoKDw/HYDDQtm1bRo4cibOzs8l8tVqN\nr6+v9NrHx4e0tDRbh1nKzz//jJeXFy1atDA7/9ixY5w9exYPDw+GDRtG69atbRzhXywd01Azx3VF\nYxpq1riuaExDzRzXFY1psH5c15nEUdtFR0djb2/PgAEDSs0bOXIkHh4eyOVyDh8+zJIlS4iIiKi2\nntMLFizAx8eHoqIiNm/ezKZNm3j55ZerJRZr/fDDD2X+Muvfvz+DBw/Gzs6Os2fP8u6777Jy5Urp\nHLxgnfLGNNSscf2wjmm4v3FdZ05V1eZCi1u2bOHmzZu89tprZv/RlPzV2adPHzQaTbX+yin+vJRK\nJSEhIVy+fNnsMqmpqdJrtVqNt7e3zWI0Jz09nV9//ZUnnnjC7HwPDw/s7Iy/tQICAvD29iY5OdmW\nIZqwdExDzRvXFY1pqFnj2pIxXbxcTRrXFY1puL9xXWcSR20ttLht2zaSkpKYOXOmdEfHvdLT06U/\nx8fHI5fL8fLyslWIJjQaDXl5eQAYDAaOHj1K8+bNSy3XvXt39u3bB8Cff/5JYmKi2btUbOnQoUME\nBgaW+Uur5Od87do1UlNTq7Ufg6VjGmrWuLZkTEPNGdeWjmmoeeO6ojEN9zeu61StqtpWaDE5OZkZ\nM2bg5+eHSqUCoF69esycOZOZM2fy5ptv4uXlxcKFC7lz5w5yuRxHR0dGjx5dbefeb926xXvvvYde\nr0ev19O4cWPGjRuHp6enScwajYaoqCiSkpKQy+WMHj2abt26VUvMxV555RXGjRtn8g+95NiIiIiQ\n4rWzs2Po0KF06dLFJrFFR0dz8uRJ7ty5g6urK66urrz//vtljul7Y6+OcW0u5tdee63MMQ1U+7g2\nF/Ps2bPLHNP3xlwd47qssQHmxzQ8+LiuU4lDEARBeHB15lSVIAiCUDlE4hAEQRCsIhKHIAiCYBWR\nOARBEASriMQhCIIgWEUkDqHOW7JkCYcOHar0ZWu6Q4cOMXfu3OoOQ6iFRMkRoVYaM2aM9OfCwkLs\n7Oykp4wnTJhQ7pOy93rrrbeqZFlrXLhwgXfeeUd6tqHY3Llzq7UeliCYIxKHUCt9/PHH0p+nTJnC\nxIkTCQgIKLWcTqdDoVDYMrT75unpybp166o7DEGokEgcwkPlwoULhIeHM2DAAL755hsCAgIYN24c\nERERXLlyBb1eT5s2bRg/frxUQ2j+/Pk88cQT9O3bl0OHDnHgwAEeeeQRfvjhB5ycnAgNDSUwMNDq\nZW/fvk1kZCRJSUk88sgj+Pn5kZeXd1/F8ebPn0/r1q05d+4cKSkpdOjQgcmTJ+Pi4gIYK6Bu27aN\n9PR0mjdvTmhoqFRJVq1Ws3nzZi5evIjBYKBnz568+OKL0ra3bNliNn5BKIu4xiE8dO7cuUNOTg5R\nUVFMnDgRg8HAk08+SVRUFFFRUahUKjZt2lTm+gkJCTRs2JBNmzbx3HPPsW7dOsoqsFDesqtXr6Zl\ny5ZER0czdOhQjhw58kD7dfjwYSZNmsT69euRy+VER0cDxi6Aq1evZuzYsWzcuJHAwECWLVuGVqtF\nr9ezbNkyfHx8iIyMZN26dfTs2fO+9lUQionEITx0ZDIZw4YNQ6lUolKpcHV1pXv37tjb2+Po6Mjg\nwYO5ePFimev7+PjQr18/5HI5ffr0ISMjg8zMTKuWVavVJCYmMnz4cOzs7Gjbtm2FLUQzMjIYO3as\nyX8ajUaa37t3b5o2bYqDgwMjRozg+PHj6PV6jh07RmBgIAEBAdjZ2fHss89SWFjI5cuXSUhIID09\nnTFjxuDg4IBKpaJt27b3ta+CUEycqhIeOm5ubiYXmQsKCvjoo4+Ij48nNzcXgPz8fPR6vdmS3h4e\nHtKf7e3tAUy+wC1ZNisrCxcXF2kaVNw8qaJrHCXLc/v4+KDT6cjKyiIjI8OkeZBcLsfHx4f09HQU\nCgW+vr5lXuexZl8FoZhIHMJDRyaTmbzevXs3KSkpLFmyBA8PD65du8asWbOq9JSMp6cnOTk5FBQU\nSF/ID9pxr2QvCrVajUKhwM3NDU9PT65fvy7NMxgMqNVqvLy8UCqVqNXqWnWTgFDziVNVwkNPo9Gg\nUqlwcnIiJyeHmJiYKn9PX19fWrZsSUxMDFqtlt9++41Tp0490DaPHDnCH3/8QUFBATt27KB79+7I\n5XJ69OjB6dOnOXfuHFqtlt27d6NUKmnTpg2tWrXC09OTrVu3otFoKCws5NKlS5W0l0JdJY44hIfe\nM888w5o1a3jxxRfx8vJi4MCB/O9//6vy9502bRpRUVH897//pVWrVvTo0UPq1mdORkaGyfMpYLzV\nuLjhUu/evYmMjCQlJYV27doxefJkABo2bMi0adOIjo6W7qqaPXu21NVt9uzZREdHM3nyZGQyGT17\n9jS5ziEI1hL9OATBRlauXEmjRo0YNmyY1euWvA1YEKqbOFUlCFUkISGBmzdvotfriY+P5+eff672\nLoeCUBnEqSpBqCJ37tzhvffeIzs7G29vb0JDQ/H396/usAThgYlTVYIgCIJVxKkqQRAEwSoicQiC\nIAhWEYlDEARBsIpIHIIgCIJVROIQBEEQrCIShyAIgmCV/wdWKY7Mg/SmFQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.65      0.40      0.50      1864\n",
            "         1.0       0.98      0.97      0.98      6567\n",
            "         2.0       0.98      0.99      0.98      1050\n",
            "         3.0       0.98      0.99      0.98       833\n",
            "         4.0       0.54      0.74      0.63      2342\n",
            "         5.0       0.88      0.87      0.88      8212\n",
            "\n",
            "    accuracy                           0.86     20868\n",
            "   macro avg       0.84      0.83      0.82     20868\n",
            "weighted avg       0.86      0.86      0.86     20868\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uPkUZnlSRlEp",
        "colab_type": "text"
      },
      "source": [
        "## Prepare the test subset ##"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5LNztCcFRbK2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df_test['x-axis'] = df_test['x-axis'] / df_test['x-axis'].max()\n",
        "df_test['y-axis'] = df_test['y-axis'] / df_test['y-axis'].max()\n",
        "df_test['z-axis'] = df_test['z-axis'] / df_test['z-axis'].max()\n",
        "\n",
        "# Round numbers\n",
        "df_test = df_test.round({'x-axis': 4, 'y-axis': 4, 'z-axis': 4})\n",
        "\n",
        "x_test, y_test = create_segments_and_labels( df_test,\n",
        "                                              TIME_PERIODS,\n",
        "                                              STEP_DISTANCE,\n",
        "                                              LABEL)\n",
        "# convert to float\n",
        "x_test = x_test.astype('float32')\n",
        "y_test = y_test.astype('float32')\n",
        "\n",
        "# convert output labels using one-hot-encoding\n",
        "y_test_hot = np_utils.to_categorical(y_test, num_classes)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "snDg3BHpTN0s",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "27ae8817-df42-4f6f-9654-8a7849502651"
      },
      "source": [
        "y_test.shape"
      ],
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(6584,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e2UPc2DrQVgD",
        "colab_type": "text"
      },
      "source": [
        "## Evaluate the model on the test set and report results with a confusion matrix ##"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iCAfeJ0FQHfg",
        "colab_type": "code",
        "outputId": "0dc7b2a7-886e-42dd-fdd3-c9709dd4a82b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "plt.figure(figsize=(6, 4))\n",
        "plt.plot(history.history['accuracy'], 'r', label='Accuracy of training data')\n",
        "plt.plot(history.history['val_accuracy'], 'b', label='Accuracy of validation data')\n",
        "plt.plot(history.history['loss'], 'r--', label='Loss of training data')\n",
        "plt.plot(history.history['val_loss'], 'b--', label='Loss of validation data')\n",
        "plt.title('Model Accuracy and Loss')\n",
        "plt.ylabel('Accuracy and Loss')\n",
        "plt.xlabel('Training Epoch')\n",
        "plt.ylim(0)\n",
        "plt.legend()\n",
        "plt.show()\n",
        "\n",
        "# Print confusion matrix for training data\n",
        "y_pred_train = model_m.predict(x_train)\n",
        "# Take the class with the highest probability from the train predictions\n",
        "max_y_pred_train = np.argmax(y_pred_train, axis=1)\n",
        "\n",
        "\n",
        "print(classification_report(y_train, max_y_pred_train))\n",
        "\n",
        "\n",
        "x_test = x_test.reshape(x_test.shape[0], input_shape)\n",
        "\n",
        "y_pred_test = model_m.predict(x_test)\n",
        "\n",
        "# Take the class with the highest probability from the test predictions\n",
        "max_y_pred_test = np.argmax(y_pred_test, axis=1)\n",
        "\n",
        "\n",
        "import seaborn as sns  # ! pip install seaborn\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import metrics\n",
        "\n",
        "def show_confusion_matrix(validations, predictions, LABELS ):\n",
        "\n",
        "    matrix = metrics.confusion_matrix(validations, predictions)\n",
        "    plt.figure(figsize=(6, 4))\n",
        "    sns.heatmap(matrix,\n",
        "                cmap='coolwarm',\n",
        "                linecolor='white',\n",
        "                linewidths=1,\n",
        "                xticklabels=LABELS,\n",
        "                yticklabels=LABELS,\n",
        "                annot=True,\n",
        "                fmt='d')\n",
        "    plt.title('Confusion Matrix')\n",
        "    plt.ylabel('True Label')\n",
        "    plt.xlabel('Predicted Label')\n",
        "    plt.show()\n",
        "\n",
        "# y_test, max_y_pred_test are (nSamples,1), LABELS is a list of the classes, indexed from 0 up to num_classes-1\n",
        "show_confusion_matrix(y_test, max_y_pred_test, LABELS )\n",
        "\n",
        "print(classification_report(y_test, max_y_pred_test))"
      ],
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEdCAYAAAAb9oCRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.65      0.40      0.50      1864\n",
            "         1.0       0.98      0.97      0.98      6567\n",
            "         2.0       0.98      0.99      0.98      1050\n",
            "         3.0       0.98      0.99      0.98       833\n",
            "         4.0       0.54      0.74      0.63      2342\n",
            "         5.0       0.88      0.87      0.88      8212\n",
            "\n",
            "    accuracy                           0.86     20868\n",
            "   macro avg       0.84      0.83      0.82     20868\n",
            "weighted avg       0.86      0.86      0.86     20868\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbMAAAFQCAYAAAA1PZCrAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.60      0.25      0.35       650\n",
            "         1.0       0.82      0.94      0.88      1990\n",
            "         2.0       0.79      1.00      0.88       452\n",
            "         3.0       0.85      0.68      0.76       370\n",
            "         4.0       0.42      0.54      0.47       725\n",
            "         5.0       0.80      0.75      0.78      2397\n",
            "\n",
            "    accuracy                           0.75      6584\n",
            "   macro avg       0.71      0.69      0.69      6584\n",
            "weighted avg       0.75      0.75      0.74      6584\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MAGQ_diISe8a",
        "colab_type": "text"
      },
      "source": [
        "# E) Change some hyperparameters and examine effect of the changes #"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x-HbYRBPSdtc",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "0450d528-fbc3-4a37-f0af-90416bcfaf30"
      },
      "source": [
        "callbacks_list = [\n",
        "    tf.keras.callbacks.ModelCheckpoint(\n",
        "        filepath='best_model.{epoch:02d}-{val_loss:.2f}.h5',\n",
        "        monitor='val_loss', save_best_only=True),\n",
        "    tf.keras.callbacks.EarlyStopping(monitor='accuracy', patience=2) # <-- changed\n",
        "]\n",
        "\n",
        "model_m.compile(loss='categorical_crossentropy',\n",
        "                optimizer='adam', metrics=['accuracy'])\n",
        "\n",
        "# Hyper-parameters\n",
        "BATCH_SIZE = 128 # <-- changed\n",
        "EPOCHS = 100  # <-- changed\n",
        "\n",
        "# Enable validation to use ModelCheckpoint and EarlyStopping callbacks.\n",
        "history = model_m.fit(x_train,\n",
        "                      y_train_hot,\n",
        "                      batch_size=BATCH_SIZE,\n",
        "                      epochs=EPOCHS,\n",
        "                      initial_epoch=0,                      \n",
        "                      callbacks=callbacks_list,\n",
        "                      validation_split=0.2,\n",
        "                      verbose=1)"
      ],
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/100\n",
            "131/131 [==============================] - 7s 55ms/step - loss: 0.3146 - accuracy: 0.8831 - val_loss: 0.7825 - val_accuracy: 0.7465\n",
            "Epoch 2/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.2931 - accuracy: 0.8900 - val_loss: 0.8241 - val_accuracy: 0.7254\n",
            "Epoch 3/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.2868 - accuracy: 0.8925 - val_loss: 0.9537 - val_accuracy: 0.6890\n",
            "Epoch 4/100\n",
            "131/131 [==============================] - 7s 54ms/step - loss: 0.2790 - accuracy: 0.8957 - val_loss: 0.9147 - val_accuracy: 0.7075\n",
            "Epoch 5/100\n",
            "131/131 [==============================] - 7s 57ms/step - loss: 0.2667 - accuracy: 0.8980 - val_loss: 0.8873 - val_accuracy: 0.7242\n",
            "Epoch 6/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.2556 - accuracy: 0.9052 - val_loss: 1.0982 - val_accuracy: 0.6423\n",
            "Epoch 7/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.2475 - accuracy: 0.9078 - val_loss: 0.8725 - val_accuracy: 0.7324\n",
            "Epoch 8/100\n",
            "131/131 [==============================] - 7s 52ms/step - loss: 0.2412 - accuracy: 0.9086 - val_loss: 0.9273 - val_accuracy: 0.7106\n",
            "Epoch 9/100\n",
            "131/131 [==============================] - 7s 52ms/step - loss: 0.2332 - accuracy: 0.9142 - val_loss: 0.9837 - val_accuracy: 0.6914\n",
            "Epoch 10/100\n",
            "131/131 [==============================] - 7s 52ms/step - loss: 0.2167 - accuracy: 0.9198 - val_loss: 1.0400 - val_accuracy: 0.6888\n",
            "Epoch 11/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.2153 - accuracy: 0.9206 - val_loss: 1.2867 - val_accuracy: 0.6114\n",
            "Epoch 12/100\n",
            "131/131 [==============================] - 7s 54ms/step - loss: 0.2074 - accuracy: 0.9236 - val_loss: 0.9637 - val_accuracy: 0.7405\n",
            "Epoch 13/100\n",
            "131/131 [==============================] - 8s 57ms/step - loss: 0.2018 - accuracy: 0.9241 - val_loss: 1.0514 - val_accuracy: 0.6981\n",
            "Epoch 14/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.1886 - accuracy: 0.9315 - val_loss: 1.1254 - val_accuracy: 0.6864\n",
            "Epoch 15/100\n",
            "131/131 [==============================] - 7s 54ms/step - loss: 0.1856 - accuracy: 0.9302 - val_loss: 1.1978 - val_accuracy: 0.6644\n",
            "Epoch 16/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.1786 - accuracy: 0.9348 - val_loss: 1.1194 - val_accuracy: 0.6768\n",
            "Epoch 17/100\n",
            "131/131 [==============================] - 7s 54ms/step - loss: 0.1787 - accuracy: 0.9341 - val_loss: 1.0703 - val_accuracy: 0.6945\n",
            "Epoch 18/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.1619 - accuracy: 0.9411 - val_loss: 1.1374 - val_accuracy: 0.7103\n",
            "Epoch 19/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.1552 - accuracy: 0.9440 - val_loss: 1.1647 - val_accuracy: 0.6885\n",
            "Epoch 20/100\n",
            "131/131 [==============================] - 7s 55ms/step - loss: 0.1485 - accuracy: 0.9469 - val_loss: 1.3915 - val_accuracy: 0.6430\n",
            "Epoch 21/100\n",
            "131/131 [==============================] - 7s 54ms/step - loss: 0.1449 - accuracy: 0.9465 - val_loss: 1.3505 - val_accuracy: 0.6691\n",
            "Epoch 22/100\n",
            "131/131 [==============================] - 7s 52ms/step - loss: 0.1365 - accuracy: 0.9517 - val_loss: 1.2687 - val_accuracy: 0.6804\n",
            "Epoch 23/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.1413 - accuracy: 0.9487 - val_loss: 1.3226 - val_accuracy: 0.6713\n",
            "Epoch 24/100\n",
            "131/131 [==============================] - 7s 55ms/step - loss: 0.1269 - accuracy: 0.9547 - val_loss: 1.2070 - val_accuracy: 0.7199\n",
            "Epoch 25/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.1210 - accuracy: 0.9554 - val_loss: 1.4233 - val_accuracy: 0.6756\n",
            "Epoch 26/100\n",
            "131/131 [==============================] - 8s 57ms/step - loss: 0.1187 - accuracy: 0.9578 - val_loss: 1.6503 - val_accuracy: 0.6198\n",
            "Epoch 27/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.1125 - accuracy: 0.9606 - val_loss: 1.3599 - val_accuracy: 0.7003\n",
            "Epoch 28/100\n",
            "131/131 [==============================] - 7s 52ms/step - loss: 0.1051 - accuracy: 0.9639 - val_loss: 1.3031 - val_accuracy: 0.7312\n",
            "Epoch 29/100\n",
            "131/131 [==============================] - 7s 52ms/step - loss: 0.0985 - accuracy: 0.9653 - val_loss: 1.5247 - val_accuracy: 0.6732\n",
            "Epoch 30/100\n",
            "131/131 [==============================] - 7s 53ms/step - loss: 0.0899 - accuracy: 0.9703 - val_loss: 1.8695 - val_accuracy: 0.6126\n",
            "Epoch 31/100\n",
            "131/131 [==============================] - 7s 52ms/step - loss: 0.0934 - accuracy: 0.9684 - val_loss: 1.4332 - val_accuracy: 0.7463\n",
            "Epoch 32/100\n",
            "131/131 [==============================] - 7s 52ms/step - loss: 0.0986 - accuracy: 0.9642 - val_loss: 1.5582 - val_accuracy: 0.6871\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9pjqn7gdS4O5",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "c87b4d81-23e5-467c-ba81-13c397dd9f6a"
      },
      "source": [
        "def draw_plot():\n",
        "  plt.figure(figsize=(6, 4))\n",
        "  plt.plot(history.history['accuracy'], 'r', label='Accuracy of training data')\n",
        "  plt.plot(history.history['val_accuracy'], 'b', label='Accuracy of validation data')\n",
        "  plt.plot(history.history['loss'], 'r--', label='Loss of training data')\n",
        "  plt.plot(history.history['val_loss'], 'b--', label='Loss of validation data')\n",
        "  plt.title('Model Accuracy and Loss')\n",
        "  plt.ylabel('Accuracy and Loss')\n",
        "  plt.xlabel('Training Epoch')\n",
        "  plt.ylim(0)\n",
        "  plt.legend()\n",
        "  plt.show()\n",
        "\n",
        "  # Print confusion matrix for training data\n",
        "  y_pred_train = model_m.predict(x_train)\n",
        "  # Take the class with the highest probability from the train predictions\n",
        "  max_y_pred_train = np.argmax(y_pred_train, axis=1)\n",
        "\n",
        "  from sklearn.metrics import classification_report\n",
        "  print(classification_report(y_train, max_y_pred_train))\n",
        "\n",
        "draw_plot()\n",
        "\n",
        "x_test = x_test.reshape(x_test.shape[0], input_shape)\n",
        "\n",
        "y_pred_test = model_m.predict(x_test)\n",
        "\n",
        "# Take the class with the highest probability from the test predictions\n",
        "max_y_pred_test = np.argmax(y_pred_test, axis=1)\n",
        "\n",
        "\n",
        "import seaborn as sns  # ! pip install seaborn\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import metrics\n",
        "\n",
        "def show_confusion_matrix(validations, predictions, LABELS ):\n",
        "\n",
        "    matrix = metrics.confusion_matrix(validations, predictions)\n",
        "    plt.figure(figsize=(6, 4))\n",
        "    sns.heatmap(matrix,\n",
        "                cmap='coolwarm',\n",
        "                linecolor='white',\n",
        "                linewidths=1,\n",
        "                xticklabels=LABELS,\n",
        "                yticklabels=LABELS,\n",
        "                annot=True,\n",
        "                fmt='d')\n",
        "    plt.title('Confusion Matrix')\n",
        "    plt.ylabel('True Label')\n",
        "    plt.xlabel('Predicted Label')\n",
        "    plt.show()\n",
        "\n",
        "# y_test, max_y_pred_test are (nSamples,1), LABELS is a list of the classes, indexed from 0 up to num_classes-1\n",
        "show_confusion_matrix(y_test, max_y_pred_test, LABELS )\n",
        "\n",
        "print(classification_report(y_test, max_y_pred_test))"
      ],
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZAAAAEdCAYAAAAikTHKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.71      0.86      0.78      1864\n",
            "         1.0       0.98      0.97      0.98      6567\n",
            "         2.0       1.00      1.00      1.00      1050\n",
            "         3.0       1.00      1.00      1.00       833\n",
            "         4.0       0.78      0.83      0.80      2342\n",
            "         5.0       0.96      0.90      0.93      8212\n",
            "\n",
            "    accuracy                           0.92     20868\n",
            "   macro avg       0.90      0.93      0.91     20868\n",
            "weighted avg       0.93      0.92      0.92     20868\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbMAAAFQCAYAAAA1PZCrAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSTETQghh86SYCSGEsHlSzIQQQti8/wfe8PasxjFOnQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.49      0.57      0.53       650\n",
            "         1.0       0.77      0.94      0.85      1990\n",
            "         2.0       0.92      1.00      0.96       452\n",
            "         3.0       0.93      0.89      0.91       370\n",
            "         4.0       0.52      0.46      0.49       725\n",
            "         5.0       0.86      0.70      0.77      2397\n",
            "\n",
            "    accuracy                           0.76      6584\n",
            "   macro avg       0.75      0.76      0.75      6584\n",
            "weighted avg       0.77      0.76      0.76      6584\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EtnNx9KKM8Mm",
        "colab_type": "text"
      },
      "source": [
        "# F) Install library ```coreML``` if not already available #\n",
        "\n",
        "...updating to module 3.1 per suggestions on [CoreML support](https://github.com/apple/coremltools/issues/446)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "crf8v96O07IS",
        "colab_type": "code",
        "outputId": "5949fc05-7057-4fe5-afbf-870f73d8801e",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        }
      },
      "source": [
        "! pip install coremltools==3.1\n",
        "\n",
        "import coremltools"
      ],
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: coremltools==3.1 in /usr/local/lib/python3.6/dist-packages (3.1)\n",
            "Requirement already satisfied: numpy>=1.14.5 in /usr/local/lib/python3.6/dist-packages (from coremltools==3.1) (1.18.2)\n",
            "Requirement already satisfied: protobuf>=3.1.0 in /usr/local/lib/python3.6/dist-packages (from coremltools==3.1) (3.10.0)\n",
            "Requirement already satisfied: six>=1.10.0 in /usr/local/lib/python3.6/dist-packages (from coremltools==3.1) (1.12.0)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.6/dist-packages (from protobuf>=3.1.0->coremltools==3.1) (46.1.3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YGWuSusjU_Ua",
        "colab_type": "text"
      },
      "source": [
        "# G) Export to CoreML file #"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V7ANwAopsEQK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "coreml_model = coremltools.converters.keras.convert(model_m,\n",
        "                                                    input_names=['acceleration'],\n",
        "                                                    output_names=['output'],\n",
        "                                                    class_labels=LABELS)\n",
        "print(coreml_model)\n",
        "coreml_model.author = 'Nils Ackermann'\n",
        "coreml_model.license = 'N/A'\n",
        "coreml_model.short_description = 'Activity based recognition based on WISDM dataset'\n",
        "coreml_model.output_description['output'] = 'Probability of each activity'\n",
        "coreml_model.output_description['classLabel'] = 'Labels of activity'\n",
        "\n",
        "coreml_model.save('HARClassifier.mlmodel')"
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}