Time_Series_Classification_Shaplet_Learning.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Time Series Classification Shaplet Learning.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyM3DIVT56FECufloz9cZv6Y",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ashishpatel26/Shapelet-time-Series-Classification/blob/main/Time_Series_Classification_Shaplet_Learning.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-s6Ife7-kPL0",
        "outputId": "312abc8c-df37-4e12-9998-17b6b7105170"
      },
      "source": [
        "!pip install tslearn"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Collecting tslearn\n",
            "\u001b[?25l  Downloading https://files.pythonhosted.org/packages/95/98/9a67e2869a8b1416eb6e6fd5e69c56f86869980403b18f30bdb5783ade9d/tslearn-0.5.0.5-cp37-cp37m-manylinux2010_x86_64.whl (790kB)\n",
            "\r\u001b[K     |▍                               | 10kB 22.8MB/s eta 0:00:01\r\u001b[K     |▉                               | 20kB 28.6MB/s eta 0:00:01\r\u001b[K     |█▎                              | 30kB 22.6MB/s eta 0:00:01\r\u001b[K     |█▋                              | 40kB 17.2MB/s eta 0:00:01\r\u001b[K     |██                              | 51kB 8.7MB/s eta 0:00:01\r\u001b[K     |██▌                             | 61kB 7.9MB/s eta 0:00:01\r\u001b[K     |███                             | 71kB 9.0MB/s eta 0:00:01\r\u001b[K     |███▎                            | 81kB 9.9MB/s eta 0:00:01\r\u001b[K     |███▊                            | 92kB 10.4MB/s eta 0:00:01\r\u001b[K     |████▏                           | 102kB 8.5MB/s eta 0:00:01\r\u001b[K     |████▋                           | 112kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████                           | 122kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████▍                          | 133kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████▉                          | 143kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████▏                         | 153kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████▋                         | 163kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████                         | 174kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████▌                        | 184kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████▉                        | 194kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████▎                       | 204kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████▊                       | 215kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████▏                      | 225kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████▌                      | 235kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████                      | 245kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████▍                     | 256kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████▊                     | 266kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████▏                    | 276kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████▋                    | 286kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████                    | 296kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████▍                   | 307kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████▉                   | 317kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████████▎                  | 327kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████████▊                  | 337kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████████                  | 348kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████████▌                 | 358kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████████                 | 368kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████████▍                | 378kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████████▊                | 389kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████████▏               | 399kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████████▋               | 409kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████████████               | 419kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████████████▍              | 430kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████████████▉              | 440kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████████████▎             | 450kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████████████▋             | 460kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████████████             | 471kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████████████▌            | 481kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████████████            | 491kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████████████▎           | 501kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████████████▊           | 512kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████████████████▏          | 522kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████████████████▌          | 532kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████████████████          | 542kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████████████████▍         | 552kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████████████████▉         | 563kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████████████████▏        | 573kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████████████████▋        | 583kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████████████████        | 593kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████████████████▌       | 604kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████████████████▉       | 614kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████████████████████▎      | 624kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████████████████████▊      | 634kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▏     | 645kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▌     | 655kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████████████████████     | 665kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▍    | 675kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▊    | 686kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████████████████████▏   | 696kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████████████████████▋   | 706kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████   | 716kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████▍  | 727kB 8.5MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████▉  | 737kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▎ | 747kB 8.5MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▊ | 757kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████ | 768kB 8.5MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████▌| 778kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 788kB 8.5MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 798kB 8.5MB/s \n",
            "\u001b[?25hRequirement already satisfied: scikit-learn in /usr/local/lib/python3.7/dist-packages (from tslearn) (0.22.2.post1)\n",
            "Requirement already satisfied: Cython in /usr/local/lib/python3.7/dist-packages (from tslearn) (0.29.22)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.7/dist-packages (from tslearn) (1.19.5)\n",
            "Requirement already satisfied: numba in /usr/local/lib/python3.7/dist-packages (from tslearn) (0.51.2)\n",
            "Requirement already satisfied: scipy in /usr/local/lib/python3.7/dist-packages (from tslearn) (1.4.1)\n",
            "Requirement already satisfied: joblib in /usr/local/lib/python3.7/dist-packages (from tslearn) (1.0.1)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.7/dist-packages (from numba->tslearn) (56.1.0)\n",
            "Requirement already satisfied: llvmlite<0.35,>=0.34.0.dev0 in /usr/local/lib/python3.7/dist-packages (from numba->tslearn) (0.34.0)\n",
            "Installing collected packages: tslearn\n",
            "Successfully installed tslearn-0.5.0.5\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FxKvTgrJjsUl",
        "outputId": "0f5831a9-08a5-4da6-e587-3936e2f0d34c"
      },
      "source": [
        "import numpy\n",
        "from sklearn.metrics import accuracy_score\n",
        "import tensorflow as tf\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "from tslearn.datasets import CachedDatasets\n",
        "from tslearn.preprocessing import TimeSeriesScalerMinMax\n",
        "from tslearn.shapelets import LearningShapelets, grabocka_params_to_shapelet_size_dict\n",
        "from tslearn.utils import ts_size"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tslearn/clustering/kmeans.py:17: UserWarning: Scikit-learn <0.24 will be deprecated in a future release of tslearn\n",
            "  \"Scikit-learn <0.24 will be deprecated in a \"\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vDd7yHU5j-3k"
      },
      "source": [
        "# Set seed for determinism\n",
        "numpy.random.seed(0)"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NF7PWz45kW3u"
      },
      "source": [
        "# Load the Trace dataset\n",
        "X_train, y_train, X_test, y_test = CachedDatasets().load_dataset(\"Trace\")"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GmsAW5Fdkbn5"
      },
      "source": [
        "# Normalize each of the timeseries in the Trace dataset\n",
        "X_train = TimeSeriesScalerMinMax().fit_transform(X_train)\n",
        "X_test = TimeSeriesScalerMinMax().fit_transform(X_test)"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hPoKPMTNmSp4"
      },
      "source": [
        "# Get statistics of the dataset\n",
        "n_ts, ts_sz = X_train.shape[:2]\n",
        "n_classes = len(set(y_train))"
      ],
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dXzUkML3nvjx",
        "outputId": "bccc1eca-7ef6-4f43-c983-8e0d371b4612"
      },
      "source": [
        "help(grabocka_params_to_shapelet_size_dict)"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Help on function grabocka_params_to_shapelet_size_dict in module tslearn.shapelets.shapelets:\n",
            "\n",
            "grabocka_params_to_shapelet_size_dict(n_ts, ts_sz, n_classes, l, r)\n",
            "    Compute number and length of shapelets.\n",
            "    \n",
            "     This function uses the heuristic from [1]_.\n",
            "    \n",
            "    Parameters\n",
            "    ----------\n",
            "    n_ts: int\n",
            "        Number of time series in the dataset\n",
            "    ts_sz: int\n",
            "        Length of time series in the dataset\n",
            "    n_classes: int\n",
            "        Number of classes in the dataset\n",
            "    l: float\n",
            "        Fraction of the length of time series to be used for base shapelet\n",
            "        length\n",
            "    r: int\n",
            "        Number of different shapelet lengths to use\n",
            "    \n",
            "    Returns\n",
            "    -------\n",
            "    dict\n",
            "        Dictionary giving, for each shapelet length, the number of such\n",
            "        shapelets to be generated\n",
            "    \n",
            "    Examples\n",
            "    --------\n",
            "    >>> d = grabocka_params_to_shapelet_size_dict(\n",
            "    ...         n_ts=100, ts_sz=100, n_classes=3, l=0.1, r=2)\n",
            "    >>> keys = sorted(d.keys())\n",
            "    >>> print(keys)\n",
            "    [10, 20]\n",
            "    >>> print([d[k] for k in keys])\n",
            "    [4, 4]\n",
            "    \n",
            "    \n",
            "    References\n",
            "    ----------\n",
            "    .. [1] J. Grabocka et al. Learning Time-Series Shapelets. SIGKDD 2014.\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RHnxcoxCmj8P",
        "outputId": "d15c28b3-6031-4cbe-970a-1e544abe3e6c"
      },
      "source": [
        "# Set the number of shapelets per size as done in the original paper\n",
        "shapelet_sizes = grabocka_params_to_shapelet_size_dict(n_ts=n_ts, ts_sz=ts_sz, n_classes=n_classes, l=0.1, r=1)\n",
        "print(shapelet_sizes)"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "{27: 4}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DDnOKomen8Ye",
        "outputId": "e8df3d11-13c8-43fe-c566-50d2a57b7b17"
      },
      "source": [
        "# Define the model using parameters provided by the authors (except that we use fewer iterations here)\n",
        "shp_clf = LearningShapelets(n_shapelets_per_size=shapelet_sizes, \n",
        "                            optimizer=tf.optimizers.Adam(.01), \n",
        "                            batch_size=16, \n",
        "                            weight_regularizer=.01, \n",
        "                            max_iter=200, \n",
        "                            random_state=42,\n",
        "                            verbose=1)"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tslearn/shapelets/shapelets.py:357: FutureWarning: The default value for 'scale' is set to False in version 0.4 to ensure backward compatibility, but is likely to change in a future version.\n",
            "  FutureWarning)\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2cet91geofzI",
        "outputId": "fab42dee-53a1-44da-d9d0-e5c2c9420a18"
      },
      "source": [
        "shp_clf.fit(X_train, y_train)"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/200\n",
            "7/7 [==============================] - 1s 3ms/step - loss: 1.4207 - categorical_accuracy: 0.3415 - categorical_crossentropy: 1.3858\n",
            "Epoch 2/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 1.3844 - categorical_accuracy: 0.3322 - categorical_crossentropy: 1.3539\n",
            "Epoch 3/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 1.3597 - categorical_accuracy: 0.3990 - categorical_crossentropy: 1.3322\n",
            "Epoch 4/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 1.3344 - categorical_accuracy: 0.5362 - categorical_crossentropy: 1.3082\n",
            "Epoch 5/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 1.3013 - categorical_accuracy: 0.5515 - categorical_crossentropy: 1.2746\n",
            "Epoch 6/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 1.2720 - categorical_accuracy: 0.5842 - categorical_crossentropy: 1.2436\n",
            "Epoch 7/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 1.2025 - categorical_accuracy: 0.6403 - categorical_crossentropy: 1.1712\n",
            "Epoch 8/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 1.2242 - categorical_accuracy: 0.5426 - categorical_crossentropy: 1.1893\n",
            "Epoch 9/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 1.1826 - categorical_accuracy: 0.5468 - categorical_crossentropy: 1.1434\n",
            "Epoch 10/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 1.1482 - categorical_accuracy: 0.5566 - categorical_crossentropy: 1.1039\n",
            "Epoch 11/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 1.1135 - categorical_accuracy: 0.5659 - categorical_crossentropy: 1.0633\n",
            "Epoch 12/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 1.0261 - categorical_accuracy: 0.6284 - categorical_crossentropy: 0.9694\n",
            "Epoch 13/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 1.0487 - categorical_accuracy: 0.5624 - categorical_crossentropy: 0.9856\n",
            "Epoch 14/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 1.0227 - categorical_accuracy: 0.5674 - categorical_crossentropy: 0.9531\n",
            "Epoch 15/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 1.0002 - categorical_accuracy: 0.5563 - categorical_crossentropy: 0.9244\n",
            "Epoch 16/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.9336 - categorical_accuracy: 0.6071 - categorical_crossentropy: 0.8517\n",
            "Epoch 17/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.9631 - categorical_accuracy: 0.5609 - categorical_crossentropy: 0.8758\n",
            "Epoch 18/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.9590 - categorical_accuracy: 0.5340 - categorical_crossentropy: 0.8658\n",
            "Epoch 19/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.8918 - categorical_accuracy: 0.5922 - categorical_crossentropy: 0.7927\n",
            "Epoch 20/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.8736 - categorical_accuracy: 0.5879 - categorical_crossentropy: 0.7694\n",
            "Epoch 21/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.8372 - categorical_accuracy: 0.6309 - categorical_crossentropy: 0.7284\n",
            "Epoch 22/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.8518 - categorical_accuracy: 0.6781 - categorical_crossentropy: 0.7387\n",
            "Epoch 23/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.8125 - categorical_accuracy: 0.8324 - categorical_crossentropy: 0.6956\n",
            "Epoch 24/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.8508 - categorical_accuracy: 0.7818 - categorical_crossentropy: 0.7314\n",
            "Epoch 25/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.7947 - categorical_accuracy: 0.9523 - categorical_crossentropy: 0.6727\n",
            "Epoch 26/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.8050 - categorical_accuracy: 0.7853 - categorical_crossentropy: 0.6801\n",
            "Epoch 27/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.7666 - categorical_accuracy: 0.7984 - categorical_crossentropy: 0.6389\n",
            "Epoch 28/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.8237 - categorical_accuracy: 0.7532 - categorical_crossentropy: 0.6924\n",
            "Epoch 29/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.7560 - categorical_accuracy: 0.8050 - categorical_crossentropy: 0.6219\n",
            "Epoch 30/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.7605 - categorical_accuracy: 0.8183 - categorical_crossentropy: 0.6240\n",
            "Epoch 31/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.7582 - categorical_accuracy: 0.9862 - categorical_crossentropy: 0.6184\n",
            "Epoch 32/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.7439 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.6009\n",
            "Epoch 33/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.7318 - categorical_accuracy: 0.8809 - categorical_crossentropy: 0.5860\n",
            "Epoch 34/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.7139 - categorical_accuracy: 0.8061 - categorical_crossentropy: 0.5655\n",
            "Epoch 35/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.7225 - categorical_accuracy: 0.8241 - categorical_crossentropy: 0.5723\n",
            "Epoch 36/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.7226 - categorical_accuracy: 0.9724 - categorical_crossentropy: 0.5695\n",
            "Epoch 37/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.7494 - categorical_accuracy: 0.7384 - categorical_crossentropy: 0.5934\n",
            "Epoch 38/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.6849 - categorical_accuracy: 0.8147 - categorical_crossentropy: 0.5270\n",
            "Epoch 39/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.7002 - categorical_accuracy: 0.8106 - categorical_crossentropy: 0.5407\n",
            "Epoch 40/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.6825 - categorical_accuracy: 0.9588 - categorical_crossentropy: 0.5206\n",
            "Epoch 41/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.6688 - categorical_accuracy: 0.8837 - categorical_crossentropy: 0.5049\n",
            "Epoch 42/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.6387 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.4738\n",
            "Epoch 43/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.6615 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.4955\n",
            "Epoch 44/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.6558 - categorical_accuracy: 0.9832 - categorical_crossentropy: 0.4881\n",
            "Epoch 45/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.6472 - categorical_accuracy: 0.7727 - categorical_crossentropy: 0.4781\n",
            "Epoch 46/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.6651 - categorical_accuracy: 0.7677 - categorical_crossentropy: 0.4947\n",
            "Epoch 47/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.6176 - categorical_accuracy: 0.8211 - categorical_crossentropy: 0.4458\n",
            "Epoch 48/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.6240 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.4514\n",
            "Epoch 49/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.6171 - categorical_accuracy: 0.9713 - categorical_crossentropy: 0.4432\n",
            "Epoch 50/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.6422 - categorical_accuracy: 0.8209 - categorical_crossentropy: 0.4672\n",
            "Epoch 51/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.6139 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.4383\n",
            "Epoch 52/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.5774 - categorical_accuracy: 0.9962 - categorical_crossentropy: 0.4013\n",
            "Epoch 53/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.6095 - categorical_accuracy: 0.8111 - categorical_crossentropy: 0.4326\n",
            "Epoch 54/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5641 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3863\n",
            "Epoch 55/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5892 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.4107\n",
            "Epoch 56/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5752 - categorical_accuracy: 0.9386 - categorical_crossentropy: 0.3956\n",
            "Epoch 57/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5749 - categorical_accuracy: 0.9351 - categorical_crossentropy: 0.3954\n",
            "Epoch 58/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5903 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.4109\n",
            "Epoch 59/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5765 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3958\n",
            "Epoch 60/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5664 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3847\n",
            "Epoch 61/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5615 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3796\n",
            "Epoch 62/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.5737 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3915\n",
            "Epoch 63/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.5438 - categorical_accuracy: 0.9889 - categorical_crossentropy: 0.3616\n",
            "Epoch 64/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.5556 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3732\n",
            "Epoch 65/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.5799 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3972\n",
            "Epoch 66/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.5495 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3666\n",
            "Epoch 67/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5477 - categorical_accuracy: 0.9924 - categorical_crossentropy: 0.3648\n",
            "Epoch 68/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5537 - categorical_accuracy: 0.7972 - categorical_crossentropy: 0.3707\n",
            "Epoch 69/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5681 - categorical_accuracy: 0.9351 - categorical_crossentropy: 0.3853\n",
            "Epoch 70/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5018 - categorical_accuracy: 0.9548 - categorical_crossentropy: 0.3191\n",
            "Epoch 71/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5824 - categorical_accuracy: 0.9351 - categorical_crossentropy: 0.3997\n",
            "Epoch 72/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5375 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3548\n",
            "Epoch 73/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5190 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3364\n",
            "Epoch 74/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.5289 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3463\n",
            "Epoch 75/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.5402 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3573\n",
            "Epoch 76/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4976 - categorical_accuracy: 0.9471 - categorical_crossentropy: 0.3148\n",
            "Epoch 77/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5303 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3478\n",
            "Epoch 78/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4798 - categorical_accuracy: 0.9962 - categorical_crossentropy: 0.2972\n",
            "Epoch 79/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.5538 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3712\n",
            "Epoch 80/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5113 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3289\n",
            "Epoch 81/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5049 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3226\n",
            "Epoch 82/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4827 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3006\n",
            "Epoch 83/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5049 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3226\n",
            "Epoch 84/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.5220 - categorical_accuracy: 0.9278 - categorical_crossentropy: 0.3394\n",
            "Epoch 85/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5001 - categorical_accuracy: 0.9784 - categorical_crossentropy: 0.3181\n",
            "Epoch 86/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.5328 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3514\n",
            "Epoch 87/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5012 - categorical_accuracy: 0.9321 - categorical_crossentropy: 0.3194\n",
            "Epoch 88/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5000 - categorical_accuracy: 0.8157 - categorical_crossentropy: 0.3187\n",
            "Epoch 89/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5038 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3235\n",
            "Epoch 90/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5012 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3204\n",
            "Epoch 91/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4884 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3072\n",
            "Epoch 92/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4838 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3032\n",
            "Epoch 93/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4904 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3095\n",
            "Epoch 94/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4929 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3119\n",
            "Epoch 95/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4843 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3035\n",
            "Epoch 96/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4652 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2845\n",
            "Epoch 97/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4711 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2902\n",
            "Epoch 98/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.5113 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3306\n",
            "Epoch 99/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4843 - categorical_accuracy: 0.9962 - categorical_crossentropy: 0.3038\n",
            "Epoch 100/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4595 - categorical_accuracy: 0.9784 - categorical_crossentropy: 0.2794\n",
            "Epoch 101/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.5012 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3210\n",
            "Epoch 102/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4768 - categorical_accuracy: 0.9687 - categorical_crossentropy: 0.2963\n",
            "Epoch 103/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4686 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2888\n",
            "Epoch 104/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4864 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3068\n",
            "Epoch 105/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4556 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2756\n",
            "Epoch 106/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4823 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3026\n",
            "Epoch 107/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4451 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2657\n",
            "Epoch 108/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4906 - categorical_accuracy: 0.9348 - categorical_crossentropy: 0.3113\n",
            "Epoch 109/200\n",
            "7/7 [==============================] - 0s 5ms/step - loss: 0.4743 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2953\n",
            "Epoch 110/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4550 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2759\n",
            "Epoch 111/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4466 - categorical_accuracy: 0.9395 - categorical_crossentropy: 0.2675\n",
            "Epoch 112/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4755 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2969\n",
            "Epoch 113/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4424 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2637\n",
            "Epoch 114/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4583 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2793\n",
            "Epoch 115/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4536 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2750\n",
            "Epoch 116/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4471 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2686\n",
            "Epoch 117/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4661 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2874\n",
            "Epoch 118/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4580 - categorical_accuracy: 0.9680 - categorical_crossentropy: 0.2792\n",
            "Epoch 119/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4322 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2540\n",
            "Epoch 120/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4849 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3067\n",
            "Epoch 121/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4456 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2671\n",
            "Epoch 122/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4594 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2812\n",
            "Epoch 123/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4635 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2854\n",
            "Epoch 124/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4633 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2851\n",
            "Epoch 125/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4532 - categorical_accuracy: 0.9771 - categorical_crossentropy: 0.2747\n",
            "Epoch 126/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4712 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2932\n",
            "Epoch 127/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4461 - categorical_accuracy: 0.9921 - categorical_crossentropy: 0.2682\n",
            "Epoch 128/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4632 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2854\n",
            "Epoch 129/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4506 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2729\n",
            "Epoch 130/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4817 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.3043\n",
            "Epoch 131/200\n",
            "7/7 [==============================] - 0s 5ms/step - loss: 0.4537 - categorical_accuracy: 0.9845 - categorical_crossentropy: 0.2765\n",
            "Epoch 132/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4642 - categorical_accuracy: 0.9074 - categorical_crossentropy: 0.2872\n",
            "Epoch 133/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4494 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2729\n",
            "Epoch 134/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4340 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2572\n",
            "Epoch 135/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4442 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2677\n",
            "Epoch 136/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4350 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2589\n",
            "Epoch 137/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4337 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2577\n",
            "Epoch 138/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4441 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2684\n",
            "Epoch 139/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4376 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2619\n",
            "Epoch 140/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4384 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2628\n",
            "Epoch 141/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4313 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2559\n",
            "Epoch 142/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4209 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2458\n",
            "Epoch 143/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4355 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2606\n",
            "Epoch 144/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4246 - categorical_accuracy: 0.9927 - categorical_crossentropy: 0.2495\n",
            "Epoch 145/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4228 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2481\n",
            "Epoch 146/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4392 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2645\n",
            "Epoch 147/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4259 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2514\n",
            "Epoch 148/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4206 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2465\n",
            "Epoch 149/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4209 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2468\n",
            "Epoch 150/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4348 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2611\n",
            "Epoch 151/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4166 - categorical_accuracy: 0.9962 - categorical_crossentropy: 0.2430\n",
            "Epoch 152/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4511 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2778\n",
            "Epoch 153/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4233 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2500\n",
            "Epoch 154/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4307 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2577\n",
            "Epoch 155/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4111 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2385\n",
            "Epoch 156/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4186 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2461\n",
            "Epoch 157/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4502 - categorical_accuracy: 0.9862 - categorical_crossentropy: 0.2776\n",
            "Epoch 158/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4014 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2292\n",
            "Epoch 159/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4147 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2423\n",
            "Epoch 160/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4217 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2496\n",
            "Epoch 161/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4317 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2598\n",
            "Epoch 162/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4148 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2429\n",
            "Epoch 163/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4171 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2452\n",
            "Epoch 164/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4405 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2686\n",
            "Epoch 165/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4204 - categorical_accuracy: 0.9757 - categorical_crossentropy: 0.2484\n",
            "Epoch 166/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4298 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2581\n",
            "Epoch 167/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4275 - categorical_accuracy: 0.9861 - categorical_crossentropy: 0.2558\n",
            "Epoch 168/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4265 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2548\n",
            "Epoch 169/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4218 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2503\n",
            "Epoch 170/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4016 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2297\n",
            "Epoch 171/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4198 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2483\n",
            "Epoch 172/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4118 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2402\n",
            "Epoch 173/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4098 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2383\n",
            "Epoch 174/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4090 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2376\n",
            "Epoch 175/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.3823 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2107\n",
            "Epoch 176/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4194 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2480\n",
            "Epoch 177/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4031 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2316\n",
            "Epoch 178/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.3990 - categorical_accuracy: 0.9927 - categorical_crossentropy: 0.2275\n",
            "Epoch 179/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.3792 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2082\n",
            "Epoch 180/200\n",
            "7/7 [==============================] - 0s 3ms/step - loss: 0.4105 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2395\n",
            "Epoch 181/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4084 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2374\n",
            "Epoch 182/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.3852 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2144\n",
            "Epoch 183/200\n",
            "7/7 [==============================] - 0s 5ms/step - loss: 0.4007 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2300\n",
            "Epoch 184/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.3879 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2175\n",
            "Epoch 185/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.3913 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2212\n",
            "Epoch 186/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.3909 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2207\n",
            "Epoch 187/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4115 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2413\n",
            "Epoch 188/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.3997 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2301\n",
            "Epoch 189/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.3820 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2124\n",
            "Epoch 190/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.3790 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2094\n",
            "Epoch 191/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4018 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2324\n",
            "Epoch 192/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.3921 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2228\n",
            "Epoch 193/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4059 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2371\n",
            "Epoch 194/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4103 - categorical_accuracy: 0.9927 - categorical_crossentropy: 0.2413\n",
            "Epoch 195/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.3591 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.1903\n",
            "Epoch 196/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4184 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2500\n",
            "Epoch 197/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4028 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2340\n",
            "Epoch 198/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.3985 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2300\n",
            "Epoch 199/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4103 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2419\n",
            "Epoch 200/200\n",
            "7/7 [==============================] - 0s 4ms/step - loss: 0.4006 - categorical_accuracy: 1.0000 - categorical_crossentropy: 0.2318\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "LearningShapelets(batch_size=16, max_iter=200, max_size=None,\n",
              "                  n_shapelets_per_size={27: 4},\n",
              "                  optimizer=<tensorflow.python.keras.optimizer_v2.adam.Adam object at 0x7f7b6aad8d50>,\n",
              "                  random_state=42, scale=False, shapelet_length=0.15,\n",
              "                  total_lengths=3, verbose=1, weight_regularizer=0.01)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rQ1hu9Ogoj3G",
        "outputId": "fc7923c6-285c-404b-d27c-de5d037519af"
      },
      "source": [
        "# Make predictions and calculate accuracy score\n",
        "pred_labels = shp_clf.predict(X_test)\n",
        "print(\"Correct classification rate:\", accuracy_score(y_test, pred_labels))"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "7/7 [==============================] - 0s 2ms/step\n",
            "Correct classification rate: 1.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 369
        },
        "id": "4SB1jSEOov8Y",
        "outputId": "3afcfba9-f42b-48ec-b1bd-cc450b2cc80f"
      },
      "source": [
        "# Plot the different discovered shapelets\n",
        "plt.figure(figsize=(14,5))\n",
        "for i, sz in enumerate(shapelet_sizes.keys()):\n",
        "    plt.subplot(len(shapelet_sizes), 1, i + 1)\n",
        "    plt.title(\"%d shapelets of size %d\" % (shapelet_sizes[sz], sz))\n",
        "    for shp in shp_clf.shapelets_:\n",
        "        if ts_size(shp) == sz:\n",
        "            plt.plot(shp.ravel())\n",
        "    plt.xlim([0, max(shapelet_sizes.keys()) - 1])\n",
        "\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1008x360 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 513
        },
        "id": "oJJPcHiyo1Nr",
        "outputId": "caca8339-b667-4f04-a5cb-2eacf41bc768"
      },
      "source": [
        "# The loss history is accessible via the `model_` that is a keras model\n",
        "plt.figure(figsize=(10,8))\n",
        "plt.plot(numpy.arange(1, shp_clf.n_iter_ + 1), shp_clf.history_[\"loss\"])\n",
        "plt.title(\"Evolution of cross-entropy loss during training\")\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.show()"
      ],
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}