multiple_Model_MFCC.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "Model_MFCC.ipynb",
      "version": "0.3.2",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Tr1a9B6pmgfv",
        "colab_type": "text"
      },
      "source": [
        "### Uploading File on Google.Colab"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "GZPj0dnxCFrF",
        "outputId": "448e3749-9b31-40d4-b3aa-4ba40b7bba84",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 124
        }
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount(\"/content/drive\")\n"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3Aietf%3Awg%3Aoauth%3A2.0%3Aoob&scope=email%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdocs.test%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive.photos.readonly%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fpeopleapi.readonly&response_type=code\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G4gBiIEUm8cG",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "05568c41-1dc1-44c7-e573-6a0210e3d4c7"
      },
      "source": [
        "cd /content/drive/My Drive"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/drive/My Drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fhsl2SrcmggK",
        "colab_type": "text"
      },
      "source": [
        "## Applying CNN on Feature Extracted\n",
        "\n",
        "#### Data Preproccessing"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "6NRi1C84C2DQ",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "from matplotlib import pyplot as plt\n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "std = StandardScaler()\n",
        "\n",
        "\n",
        "data = pd.DataFrame(np.load(\"dataset.npy\",allow_pickle= True))\n",
        "data.columns = ['feature', 'label']\n",
        "X = np.array(data.feature.tolist())\n",
        "std.fit(X)\n",
        "Y = np.array(data.label.tolist())\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4dg-XwYCr-ou",
        "colab_type": "text"
      },
      "source": [
        "**Ploting Distribution of Data Set**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Tumw1DhgqsNs",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 305
        },
        "outputId": "ee3050e0-6a9d-4ead-ab7a-0ba5274f296c"
      },
      "source": [
        "%matplotlib inline\n",
        "plt.hist(Y, bins = 20);\n",
        "print(np.unique(Y))"
      ],
      "execution_count": 209,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "['air_conditioner' 'car_horn' 'children_playing' 'dog_bark' 'drilling'\n",
            " 'engine_idling' 'gun_shot' 'jackhammer' 'siren' 'street_music']\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYAAAAD9CAYAAAC1DKAUAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nzCrlADAzq5QDwMysUg4AM7NK/X/yXgLSDHY3YgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iN13JJGwz7id",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "e883d65f-c5f0-4014-b2f6-6c99d4f7ec12"
      },
      "source": [
        "X.shape,Y.shape,X[0].reshape(1,X[0].shape[0]).shape"
      ],
      "execution_count": 210,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((8732, 40), (8732,), (1, 40))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 210
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-AfmgWk8zqgM",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "X,val_x,y,val_y = train_test_split(X,Y,test_size=0.3)\n",
        "lb = LabelEncoder()\n",
        "\n",
        "from keras.utils import np_utils\n",
        "\n",
        "y_en = np_utils.to_categorical(lb.fit_transform(y))\n",
        "val_y_en = np_utils.to_categorical(lb.fit_transform(val_y))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0WcVU3JJqcb5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def acc(y_test,prediction):\n",
        "\n",
        "    ### PRINTING ACCURACY OF PREDICTION\n",
        "    ### RECALL\n",
        "    ### PRECISION\n",
        "    ### CLASIFICATION REPORT\n",
        "    ### CONFUSION MATRIX\n",
        "    cm = confusion_matrix(y_test, prediction)\n",
        "    recall = np.diag(cm) / np.sum(cm, axis = 1)\n",
        "    precision = np.diag(cm) / np.sum(cm, axis = 0)\n",
        "    \n",
        "    print ('Recall:', recall)\n",
        "    print ('Precision:', precision)\n",
        "    print ('\\n clasification report:\\n', classification_report(y_test,prediction))\n",
        "    print ('\\n confussion matrix:\\n',confusion_matrix(y_test, prediction))\n",
        "    \n",
        "    ax = sns.heatmap(confusion_matrix(y_test, prediction),linewidths= 0.5,cmap=\"YlGnBu\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oPcdhRyJmggU",
        "colab_type": "text"
      },
      "source": [
        "#### Model Structure\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "HYeTNd3Jcepb",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Dropout, Activation, Flatten\n",
        "from keras.layers import Convolution2D, MaxPooling2D\n",
        "from keras.optimizers import Adam\n",
        "from keras.utils import np_utils\n",
        "from sklearn import metrics \n",
        "from sklearn.metrics import precision_score, recall_score, confusion_matrix, classification_report, accuracy_score, f1_score\n",
        "from keras.callbacks import LearningRateScheduler\n",
        "\n",
        "\n",
        "\n",
        "num_labels = y_en.shape[1]\n",
        "nets = 5\n",
        "\n",
        "model = [0] *nets\n",
        "\n",
        "# build model\n",
        "for net in range(nets):\n",
        "  model[net] = Sequential()\n",
        "\n",
        "\n",
        "  model[net].add(Dense(512, input_shape=(40,)))\n",
        "  model[net].add(Activation('relu'))\n",
        "  model[net].add(Dropout(0.45))\n",
        "\n",
        "\n",
        "  model[net].add(Dense(256))\n",
        "  model[net].add(Activation('relu'))\n",
        "  model[net].add(Dropout(0.45))\n",
        "\n",
        "\n",
        "  model[net].add(Dense(num_labels))\n",
        "  model[net].add(Activation('softmax'))\n",
        "\n",
        "\n",
        "\n",
        "  model[net].compile(loss='categorical_crossentropy', metrics=['accuracy'], optimizer='RMSprop')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "z9IQhRkMmggh",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 416
        },
        "outputId": "4b64659d-747b-45d7-8209-384394f603a3"
      },
      "source": [
        "model[-1].summary()"
      ],
      "execution_count": 229,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_424 (Dense)            (None, 512)               20992     \n",
            "_________________________________________________________________\n",
            "activation_424 (Activation)  (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "dropout_295 (Dropout)        (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "dense_425 (Dense)            (None, 256)               131328    \n",
            "_________________________________________________________________\n",
            "activation_425 (Activation)  (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dropout_296 (Dropout)        (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dense_426 (Dense)            (None, 10)                2570      \n",
            "_________________________________________________________________\n",
            "activation_426 (Activation)  (None, 10)                0         \n",
            "=================================================================\n",
            "Total params: 154,890\n",
            "Trainable params: 154,890\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "EixAes_Hcl-A",
        "outputId": "ae8503a2-9520-41e0-ffd2-c0b4ce49425a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 104
        }
      },
      "source": [
        "\n",
        "\n",
        "\n",
        "# TRAIN NETWORKS\n",
        "history = [0] * nets\n",
        "epochs = 132\n",
        "for j in range(nets):\n",
        "    X_train2, X_val2, Y_train2, Y_val2 = X,val_x, y_en, val_y_en\n",
        "    history[j] = model[j].fit(X,Y_train2, batch_size=256,\n",
        "        epochs = epochs,   \n",
        "        validation_data = (X_val2,Y_val2),  verbose=0)\n",
        "    print(\"CNN {0:d}: Epochs={1:d}, Train accuracy={2:.5f}, Validation accuracy={3:.5f}\".format(\n",
        "        j+1,epochs,max(history[j].history['acc']),max(history[j].history['val_acc']) ))\n",
        "\n"
      ],
      "execution_count": 231,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CNN 1: Epochs=132, Train accuracy=0.92539, Validation accuracy=0.91527\n",
            "CNN 2: Epochs=132, Train accuracy=0.93063, Validation accuracy=0.91527\n",
            "CNN 3: Epochs=132, Train accuracy=0.92490, Validation accuracy=0.91794\n",
            "CNN 4: Epochs=132, Train accuracy=0.92736, Validation accuracy=0.92214\n",
            "CNN 5: Epochs=132, Train accuracy=0.92408, Validation accuracy=0.91947\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "CimDBPi7Xldp",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 718
        },
        "outputId": "1aa74623-6c99-41f9-ec66-055abeddc5f2"
      },
      "source": [
        "net = -1\n",
        "name_title = ['Loss','Accuracy']\n",
        "fig=plt.figure(figsize=(64,64))\n",
        "for i in range(0,2):\n",
        "    ax=fig.add_subplot(8,8,i+1)\n",
        "    plt.plot(history[net].history[list(history[net].history.keys())[i]], label = list(history[net].history.keys())[i] )\n",
        "    plt.plot(history[net].history[list(history[net].history.keys())[i+2]],label = list(history[net].history.keys())[i+2] )\n",
        "    plt.xlabel('Epochs', fontsize=18)\n",
        "    plt.ylabel(name_title[i], fontsize=18)\n",
        "    plt.legend()\n",
        "    plt.show()"
      ],
      "execution_count": 232,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAbcAAAGrCAYAAABDg9kLAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nchRuIiISOQo3ERGJnP8PHR4WIMVk/18AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 4608x4608 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY0AAAESCAYAAAABl4lHAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/r+9OhAAAAAAEORvvcIAJREAmzQA2ALZo73nbdiXOAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IzYKs0ar44vp",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 877
        },
        "outputId": "14e82e32-913c-4649-d3da-c09c5afc95d2"
      },
      "source": [
        "results = np.zeros( (val_x.shape[0],10) ) \n",
        "for j in range(nets):\n",
        "  results = results  + model[j].predict(val_x)\n",
        "results = np.argmax(results,axis = 1)\n",
        "val_y_n = np.argmax(val_y_en,axis =1)\n",
        "acc(val_y_n,results)"
      ],
      "execution_count": 233,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Recall: [0.97697368 0.92307692 0.91467577 0.83501684 0.93831169 0.97333333\n",
            " 0.85483871 0.97526502 0.95255474 0.85034014]\n",
            "Precision: [0.95806452 0.95652174 0.80722892 0.91176471 0.96655518 0.92993631\n",
            " 0.9137931  0.95172414 0.95255474 0.90909091]\n",
            "\n",
            " clasification report:\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.96      0.98      0.97       304\n",
            "           1       0.96      0.92      0.94       143\n",
            "           2       0.81      0.91      0.86       293\n",
            "           3       0.91      0.84      0.87       297\n",
            "           4       0.97      0.94      0.95       308\n",
            "           5       0.93      0.97      0.95       300\n",
            "           6       0.91      0.85      0.88       124\n",
            "           7       0.95      0.98      0.96       283\n",
            "           8       0.95      0.95      0.95       274\n",
            "           9       0.91      0.85      0.88       294\n",
            "\n",
            "    accuracy                           0.92      2620\n",
            "   macro avg       0.93      0.92      0.92      2620\n",
            "weighted avg       0.92      0.92      0.92      2620\n",
            "\n",
            "\n",
            " confussion matrix:\n",
            " [[297   0   2   0   0   0   1   1   1   2]\n",
            " [  0 132   0   5   1   0   0   1   2   2]\n",
            " [  3   2 268   4   0   3   5   0   1   7]\n",
            " [  2   3  18 248   4   7   4   0   4   7]\n",
            " [  1   0   6   2 289   0   0   9   0   1]\n",
            " [  0   0   7   0   0 292   0   0   0   1]\n",
            " [  0   1   6   5   0   1 106   1   2   2]\n",
            " [  0   0   0   0   4   1   0 276   0   2]\n",
            " [  0   0   8   2   0   2   0   0 261   1]\n",
            " [  7   0  17   6   1   8   0   2   3 250]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWAAAAD8CAYAAABJsn7AAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "m8M_7fEdXm10"
      },
      "source": [
        "## **Implementing XGBOOST Algorithm on same Dataset**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Js9E123sKekZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from xgboost import XGBClassifier"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "RZkY89khHeF6",
        "colab": {}
      },
      "source": [
        "train = np.argmax(y_en,axis =  1)\n",
        "test = np.argmax(val_y_en,axis =  1)\n",
        "test_x = val_x"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "BbPVx1WcTY_J",
        "colab": {}
      },
      "source": [
        "model_xgb = XGBClassifier(learning_rate =0.1, \n",
        "                      n_estimators=400, \n",
        "                      max_depth=8,\n",
        "                      objective= 'multi:softmax', \n",
        "                      nthread=4, )\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "W87KEg2JWJrA",
        "outputId": "a4ef3561-34be-425a-b96d-a656d8162b86",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 139
        }
      },
      "source": [
        "model_xgb.fit(X,train)"
      ],
      "execution_count": 222,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "XGBClassifier(base_score=0.5, booster='gbtree', colsample_bylevel=1,\n",
              "              colsample_bynode=1, colsample_bytree=1, gamma=0,\n",
              "              learning_rate=0.1, max_delta_step=0, max_depth=8,\n",
              "              min_child_weight=1, missing=None, n_estimators=400, n_jobs=1,\n",
              "              nthread=4, objective='multi:softprob', random_state=0,\n",
              "              reg_alpha=0, reg_lambda=1, scale_pos_weight=1, seed=None,\n",
              "              silent=None, subsample=1, verbosity=1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 222
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "aFOGDEl8WOOl",
        "colab": {}
      },
      "source": [
        "pred = model_xgb.predict(test_x)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "Aikv-HFZXfYU",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 877
        },
        "outputId": "4d2c3a8d-36a3-4f2e-ebba-d33e47d5c125"
      },
      "source": [
        "acc(test,pred)"
      ],
      "execution_count": 224,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Recall: [0.95394737 0.85314685 0.86006826 0.7979798  0.88636364 0.94666667\n",
            " 0.79032258 0.96466431 0.91970803 0.84353741]\n",
            "Precision: [0.95394737 0.93846154 0.7826087  0.86181818 0.90397351 0.92508143\n",
            " 0.96078431 0.90397351 0.94382022 0.802589  ]\n",
            "\n",
            " clasification report:\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.95      0.95      0.95       304\n",
            "           1       0.94      0.85      0.89       143\n",
            "           2       0.78      0.86      0.82       293\n",
            "           3       0.86      0.80      0.83       297\n",
            "           4       0.90      0.89      0.90       308\n",
            "           5       0.93      0.95      0.94       300\n",
            "           6       0.96      0.79      0.87       124\n",
            "           7       0.90      0.96      0.93       283\n",
            "           8       0.94      0.92      0.93       274\n",
            "           9       0.80      0.84      0.82       294\n",
            "\n",
            "    accuracy                           0.89      2620\n",
            "   macro avg       0.90      0.88      0.89      2620\n",
            "weighted avg       0.89      0.89      0.89      2620\n",
            "\n",
            "\n",
            " confussion matrix:\n",
            " [[290   0   2   0   1   1   0   4   0   6]\n",
            " [  0 122   2   1   5   5   0   2   2   4]\n",
            " [  1   3 252   5   4   3   2   5   2  16]\n",
            " [  5   2  23 237   5   4   1   1   4  15]\n",
            " [  1   0   6   5 273   0   0  12   1  10]\n",
            " [  1   1   4   0   2 284   0   1   3   4]\n",
            " [  0   0   9  10   2   1  98   1   0   3]\n",
            " [  1   0   2   0   4   0   0 273   0   3]\n",
            " [  0   1   7   6   0   6   1   1 252   0]\n",
            " [  5   1  15  11   6   3   0   2   3 248]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWAAAAD8CAYAAABJsn7AAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "fPBB7VWQiE5i",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 877
        },
        "outputId": "7963bc42-3757-44d3-9764-edf276e50ac0"
      },
      "source": [
        "results = np.zeros( (val_x.shape[0],10) ) \n",
        "for j in range(nets):\n",
        "  results = results  + model[j].predict_proba(val_x) + model_xgb.predict_proba(val_x)\n",
        "results = np.argmax(results,axis = 1)\n",
        "acc(test,results)"
      ],
      "execution_count": 225,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Recall: [0.97368421 0.9020979  0.92150171 0.84848485 0.92857143 0.95666667\n",
            " 0.82258065 0.97526502 0.95620438 0.8707483 ]\n",
            "Precision: [0.96732026 0.97727273 0.81570997 0.90322581 0.94701987 0.94098361\n",
            " 0.97142857 0.93877551 0.95970696 0.87372014]\n",
            "\n",
            " clasification report:\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.97      0.97      0.97       304\n",
            "           1       0.98      0.90      0.94       143\n",
            "           2       0.82      0.92      0.87       293\n",
            "           3       0.90      0.85      0.88       297\n",
            "           4       0.95      0.93      0.94       308\n",
            "           5       0.94      0.96      0.95       300\n",
            "           6       0.97      0.82      0.89       124\n",
            "           7       0.94      0.98      0.96       283\n",
            "           8       0.96      0.96      0.96       274\n",
            "           9       0.87      0.87      0.87       294\n",
            "\n",
            "    accuracy                           0.92      2620\n",
            "   macro avg       0.93      0.92      0.92      2620\n",
            "weighted avg       0.92      0.92      0.92      2620\n",
            "\n",
            "\n",
            " confussion matrix:\n",
            " [[296   0   1   0   1   1   0   1   0   4]\n",
            " [  0 129   1   3   1   1   0   2   2   4]\n",
            " [  2   1 270   2   1   2   1   1   1  12]\n",
            " [  2   1  20 252   4   6   2   0   3   7]\n",
            " [  1   0   4   3 286   0   0  11   0   3]\n",
            " [  0   1   6   0   1 287   0   0   3   2]\n",
            " [  1   0   9   7   1   0 102   1   0   3]\n",
            " [  0   0   1   0   4   0   0 276   0   2]\n",
            " [  0   0   5   3   0   3   0   1 262   0]\n",
            " [  4   0  14   9   3   5   0   1   2 256]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWAAAAD8CAYAAABJsn7AAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nG3hp3qYAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qK7dYcWaJWP1",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 877
        },
        "outputId": "0f60985e-f1bc-4730-bac0-1f23f0d30d7c"
      },
      "source": [
        "results = np.zeros( (val_x.shape[0],10) ) \n",
        "for j in range(nets):\n",
        "  results = results  + model[j].predict_proba(val_x) + nets*model_xgb.predict_proba(val_x)\n",
        "results = np.argmax(results,axis = 1)\n",
        "acc(test,results)"
      ],
      "execution_count": 227,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Recall: [0.95394737 0.88111888 0.87372014 0.80808081 0.8961039  0.94666667\n",
            " 0.79032258 0.97173145 0.91970803 0.86054422]\n",
            "Precision: [0.96345515 0.94736842 0.79012346 0.86330935 0.91390728 0.93114754\n",
            " 0.97029703 0.90460526 0.9509434  0.82410423]\n",
            "\n",
            " clasification report:\n",
            "               precision    recall  f1-score   support\n",
            "\n",
            "           0       0.96      0.95      0.96       304\n",
            "           1       0.95      0.88      0.91       143\n",
            "           2       0.79      0.87      0.83       293\n",
            "           3       0.86      0.81      0.83       297\n",
            "           4       0.91      0.90      0.90       308\n",
            "           5       0.93      0.95      0.94       300\n",
            "           6       0.97      0.79      0.87       124\n",
            "           7       0.90      0.97      0.94       283\n",
            "           8       0.95      0.92      0.94       274\n",
            "           9       0.82      0.86      0.84       294\n",
            "\n",
            "    accuracy                           0.90      2620\n",
            "   macro avg       0.91      0.89      0.90      2620\n",
            "weighted avg       0.90      0.90      0.90      2620\n",
            "\n",
            "\n",
            " confussion matrix:\n",
            " [[290   0   2   0   1   1   0   4   0   6]\n",
            " [  0 126   2   2   3   2   0   2   2   4]\n",
            " [  1   3 256   5   4   3   1   4   1  15]\n",
            " [  4   2  23 240   5   4   1   1   3  14]\n",
            " [  1   0   6   5 276   0   0  13   1   6]\n",
            " [  1   1   4   0   2 284   0   1   3   4]\n",
            " [  0   0  10   9   2   1  98   1   0   3]\n",
            " [  0   0   1   0   5   0   0 275   0   2]\n",
            " [  0   1   7   6   0   6   1   1 252   0]\n",
            " [  4   0  13  11   4   4   0   2   3 253]]\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWAAAAD8CAYAAABJsn7AAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nm5kVxAnYzKwgTsBmZgVxAjYzK8j/A7+KlPakVyngAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "l_m61yzgJbzr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}