Mega_Case_Study.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Mega_Case_Study.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/Shakilgithub20/Basic-Computer-Vision/blob/main/Mega_Case_Study.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GJYk9gRO_W9P"
      },
      "source": [
        "#Mega Case Study - Making a hybrid Deep Learning Model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D8PA8C_X_fHe"
      },
      "source": [
        "#Part 1 - SOM"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "48btSZ8cAAVt"
      },
      "source": [
        "##Install MiniSom Package"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "outputId": "3530328f-8332-45b2-cadc-dec1cd573540",
        "id": "5ueD5zlVAAVu",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "!pip install MiniSom"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: MiniSom in /usr/local/lib/python3.6/dist-packages (2.2.5)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "y2L7M7QLAAVx"
      },
      "source": [
        "## Importing the libraries\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zy0_GMcbAAVx"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M74YAQUJAAVz"
      },
      "source": [
        "## Importing the dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oP387iJvAAV0"
      },
      "source": [
        "dataset = pd.read_csv('Credit_Card_Applications.csv')\n",
        "X = dataset.iloc[:, :-1].values \n",
        "y = dataset.iloc[:, -1].values"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xaNRHZwYAAV2"
      },
      "source": [
        "## Feature Scaling\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "57QK6WXTAAV3"
      },
      "source": [
        "from sklearn.preprocessing import MinMaxScaler\n",
        "sc = MinMaxScaler(feature_range = (0,1))\n",
        "X = sc.fit_transform(X)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WyD5sEHYAAV5"
      },
      "source": [
        "##Training the SOM\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-BXn8YphAAV6"
      },
      "source": [
        "from minisom import MiniSom\n",
        "som = MiniSom(x=10, y=10, input_len= 15, sigma= 1.0, learning_rate = 0.5)\n",
        "som.random_weights_init(X)\n",
        "som.train_random(data = X, num_iteration = 100)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "d9qf8w7sAAV8"
      },
      "source": [
        "##Visualizing the results\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cshl06SmAAV8",
        "outputId": "e183b980-8d59-473a-d72d-f3086473e90d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 269
        }
      },
      "source": [
        "from pylab import bone, pcolor, colorbar, plot, show\n",
        "bone()\n",
        "pcolor(som.distance_map().T)\n",
        "colorbar()\n",
        "markers = ['o', 's']\n",
        "colors = ['r', 'g']\n",
        "for i, x in enumerate(X):\n",
        "    w = som.winner(x)\n",
        "    plot(w[0] + 0.5,\n",
        "         w[1] + 0.5,\n",
        "         markers[y[i]],\n",
        "         markeredgecolor = colors[y[i]],\n",
        "         markerfacecolor = 'None',\n",
        "         markersize = 10,\n",
        "         markeredgewidth = 2)\n",
        "show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gijZkcIGAAWA"
      },
      "source": [
        "## Finding the frauds\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EJIH9sDvAAWA"
      },
      "source": [
        "mappings = som.win_map(X)\n",
        "frauds = np.concatenate((mappings[(6,8)], mappings[(5,1)]), axis = 0)\n",
        "frauds = sc.inverse_transform(frauds)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EOVO2DzCAAWD"
      },
      "source": [
        "##Printing the Fraunch Clients"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "outputId": "6babd385-f60f-42db-b5cf-c7b881da768f",
        "id": "IjotUwR3AAWD",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 833
        }
      },
      "source": [
        "print('Fraud Customer IDs')\n",
        "for i in frauds[:, 0]:\n",
        "  print(int(i))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Fraud Customer IDs\n",
            "15644446\n",
            "15803378\n",
            "15714087\n",
            "15730793\n",
            "15775750\n",
            "15788776\n",
            "15765257\n",
            "15654859\n",
            "15723827\n",
            "15767231\n",
            "15746995\n",
            "15762716\n",
            "15706464\n",
            "15725028\n",
            "15768474\n",
            "15764444\n",
            "15762045\n",
            "15731989\n",
            "15671534\n",
            "15721569\n",
            "15799859\n",
            "15771409\n",
            "15777893\n",
            "15674750\n",
            "15667460\n",
            "15684512\n",
            "15805212\n",
            "15691150\n",
            "15785339\n",
            "15740458\n",
            "15807546\n",
            "15773869\n",
            "15796111\n",
            "15809800\n",
            "15802071\n",
            "15790630\n",
            "15750055\n",
            "15748499\n",
            "15803682\n",
            "15770406\n",
            "15732943\n",
            "15770255\n",
            "15773789\n",
            "15794868\n",
            "15762392\n",
            "15812470\n",
            "15674583\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-9oatjNmAC8i"
      },
      "source": [
        "#Part 2 - Going from Unsupervised to Supervised Deep Learning"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eDv-ecVtAtwU"
      },
      "source": [
        "##Create Matrix of Features"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h3ADq7HsA3Sh"
      },
      "source": [
        "customers = dataset.iloc[:, 1:].values"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ycrQvNAvBAng"
      },
      "source": [
        "## Create Dependent Variable"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Dt1TQG3NBFDM"
      },
      "source": [
        "is_fraud = np.zeros(len(dataset))\n",
        "for i in range(len(dataset)):\n",
        "  if dataset.iloc[i,0] in frauds:\n",
        "    is_fraud[i] = 1"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OgQBLi0iBsji"
      },
      "source": [
        "#Part 3 - ANN"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RE_FcHyfV3TQ"
      },
      "source": [
        "### Feature Scaling"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ViCrE00rV8Sk"
      },
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "sc = StandardScaler()\n",
        "customers = sc.fit_transform(customers)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-zfEzkRVXIwF"
      },
      "source": [
        "## Building the ANN"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S5-uTVVvCcTf"
      },
      "source": [
        "##Import Tensorflow"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "veBa6HoECbZI",
        "outputId": "bec46426-fda2-40fd-832c-09b6a44001a9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "import tensorflow as tf\n",
        "tf.__version__"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'2.2.0'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KvdeScabXtlB"
      },
      "source": [
        "## Initializing the ANN"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3dtrScHxXQox"
      },
      "source": [
        "ann = tf.keras.models.Sequential()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rP6urV6SX7kS"
      },
      "source": [
        "##Adding the input layer and the first hidden layer"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bppGycBXYCQr"
      },
      "source": [
        "ann.add(tf.keras.layers.Dense(units=2, activation='relu'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OyNEe6RXYcU4"
      },
      "source": [
        "## Adding the output layer"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Cn3x41RBYfvY"
      },
      "source": [
        "ann.add(tf.keras.layers.Dense(units=1, activation='sigmoid'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JT4u2S1_Y4WG"
      },
      "source": [
        "## Training the ANN"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8GWlJChhY_ZI"
      },
      "source": [
        "## Compiling the ANN"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fG3RrwDXZEaS"
      },
      "source": [
        "ann.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0QR_G5u7ZLSM"
      },
      "source": [
        "## Training the ANN on the Training set"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nHZ-LKv_ZRb3",
        "outputId": "b9f6c5ec-47ae-4ebf-868b-d72ed3c02996",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 374
        }
      },
      "source": [
        "ann.fit(customers, is_fraud, batch_size = 1, epochs = 10)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "690/690 [==============================] - 1s 921us/step - loss: 1.0211 - accuracy: 0.4449\n",
            "Epoch 2/10\n",
            "690/690 [==============================] - 1s 943us/step - loss: 0.5365 - accuracy: 0.8536\n",
            "Epoch 3/10\n",
            "690/690 [==============================] - 1s 920us/step - loss: 0.3965 - accuracy: 0.9232\n",
            "Epoch 4/10\n",
            "690/690 [==============================] - 1s 911us/step - loss: 0.3335 - accuracy: 0.9290\n",
            "Epoch 5/10\n",
            "690/690 [==============================] - 1s 924us/step - loss: 0.2987 - accuracy: 0.9304\n",
            "Epoch 6/10\n",
            "690/690 [==============================] - 1s 908us/step - loss: 0.2786 - accuracy: 0.9304\n",
            "Epoch 7/10\n",
            "690/690 [==============================] - 1s 945us/step - loss: 0.2662 - accuracy: 0.9304\n",
            "Epoch 8/10\n",
            "690/690 [==============================] - 1s 920us/step - loss: 0.2581 - accuracy: 0.9304\n",
            "Epoch 9/10\n",
            "690/690 [==============================] - 1s 904us/step - loss: 0.2509 - accuracy: 0.9304\n",
            "Epoch 10/10\n",
            "690/690 [==============================] - 1s 908us/step - loss: 0.2408 - accuracy: 0.9304\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tensorflow.python.keras.callbacks.History at 0x7f50404266a0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tJj5k2MxZga3"
      },
      "source": [
        "## Predicting test set results"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nIyEeQdRZwgs"
      },
      "source": [
        "y_pred = ann.predict(customers)\n",
        "y_pred = np.concatenate((dataset.iloc[:, 0:1].values, y_pred), axis = 1)\n",
        "y_pred = y_pred[y_pred[:, 1].argsort()]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fX2iDXYiD-sN"
      },
      "source": [
        "print(y_pred)"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}