Poisson_Regression.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# Poisson Regression"
      ],
      "metadata": {}
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n",
        "\n",
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")\n",
        "\n",
        "import yfinance as yf\n",
        "yf.pdr_override()"
      ],
      "outputs": [],
      "execution_count": 1,
      "metadata": {
        "collapsed": false,
        "outputHidden": false,
        "inputHidden": false,
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:44.313Z",
          "iopub.execute_input": "2020-05-24T05:05:44.317Z",
          "iopub.status.idle": "2020-05-24T05:05:45.260Z",
          "shell.execute_reply": "2020-05-24T05:05:45.268Z"
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# input\n",
        "symbol = 'AMD'\n",
        "start = '2014-01-01'\n",
        "end = '2018-08-27'\n",
        "\n",
        "# Read data \n",
        "dataset = yf.download(symbol,start,end)\n",
        "\n",
        "# View Columns\n",
        "dataset.head()"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[*********************100%***********************]  1 of 1 completed\n"
          ]
        },
        {
          "output_type": "execute_result",
          "execution_count": 2,
          "data": {
            "text/plain": [
              "            Adj Close  Close  High   Low  Open    Volume\n",
              "Date                                                    \n",
              "2014-01-02       3.95   3.95  3.98  3.84  3.85  20548400\n",
              "2014-01-03       4.00   4.00  4.00  3.88  3.98  22887200\n",
              "2014-01-06       4.13   4.13  4.18  3.99  4.01  42398300\n",
              "2014-01-07       4.18   4.18  4.25  4.11  4.19  42932100\n",
              "2014-01-08       4.18   4.18  4.26  4.14  4.23  30678700"
            ],
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Adj Close</th>\n",
              "      <th>Close</th>\n",
              "      <th>High</th>\n",
              "      <th>Low</th>\n",
              "      <th>Open</th>\n",
              "      <th>Volume</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Date</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2014-01-02</th>\n",
              "      <td>3.95</td>\n",
              "      <td>3.95</td>\n",
              "      <td>3.98</td>\n",
              "      <td>3.84</td>\n",
              "      <td>3.85</td>\n",
              "      <td>20548400</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2014-01-03</th>\n",
              "      <td>4.00</td>\n",
              "      <td>4.00</td>\n",
              "      <td>4.00</td>\n",
              "      <td>3.88</td>\n",
              "      <td>3.98</td>\n",
              "      <td>22887200</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2014-01-06</th>\n",
              "      <td>4.13</td>\n",
              "      <td>4.13</td>\n",
              "      <td>4.18</td>\n",
              "      <td>3.99</td>\n",
              "      <td>4.01</td>\n",
              "      <td>42398300</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2014-01-07</th>\n",
              "      <td>4.18</td>\n",
              "      <td>4.18</td>\n",
              "      <td>4.25</td>\n",
              "      <td>4.11</td>\n",
              "      <td>4.19</td>\n",
              "      <td>42932100</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2014-01-08</th>\n",
              "      <td>4.18</td>\n",
              "      <td>4.18</td>\n",
              "      <td>4.26</td>\n",
              "      <td>4.14</td>\n",
              "      <td>4.23</td>\n",
              "      <td>30678700</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ]
          },
          "metadata": {}
        }
      ],
      "execution_count": 2,
      "metadata": {
        "collapsed": false,
        "outputHidden": false,
        "inputHidden": false,
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:45.275Z",
          "iopub.execute_input": "2020-05-24T05:05:45.279Z",
          "iopub.status.idle": "2020-05-24T05:05:46.647Z",
          "shell.execute_reply": "2020-05-24T05:05:46.693Z"
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "dataset['Increase_Decrease'] = np.where(dataset['Volume'].shift(-1) > dataset['Volume'],1,0)\n",
        "dataset['Buy_Sell_on_Open'] = np.where(dataset['Open'].shift(-1) > dataset['Open'],1,0)\n",
        "dataset['Buy_Sell'] = np.where(dataset['Adj Close'].shift(-1) > dataset['Adj Close'],1,0)\n",
        "dataset['Returns'] = dataset['Adj Close'].pct_change()\n",
        "dataset = dataset.dropna()"
      ],
      "outputs": [],
      "execution_count": 3,
      "metadata": {
        "collapsed": true,
        "outputExpanded": false,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:46.658Z",
          "iopub.execute_input": "2020-05-24T05:05:46.664Z",
          "iopub.status.idle": "2020-05-24T05:05:46.671Z",
          "shell.execute_reply": "2020-05-24T05:05:46.696Z"
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "dataset.tail()"
      ],
      "outputs": [
        {
          "output_type": "execute_result",
          "execution_count": 4,
          "data": {
            "text/plain": [
              "            Adj Close      Close   High        Low       Open     Volume  \\\n",
              "Date                                                                       \n",
              "2018-08-20  19.980000  19.980000  20.08  19.350000  19.790001   62983200   \n",
              "2018-08-21  20.400000  20.400000  20.42  19.860001  19.980000   55629000   \n",
              "2018-08-22  20.900000  20.900000  20.92  20.209999  20.280001   62002700   \n",
              "2018-08-23  22.290001  22.290001  22.32  21.139999  21.190001  113444100   \n",
              "2018-08-24  23.980000  23.980000  24.00  22.670000  22.910000  164328200   \n",
              "\n",
              "            Increase_Decrease  Buy_Sell_on_Open  Buy_Sell   Returns  \n",
              "Date                                                                 \n",
              "2018-08-20                  0                 1         1  0.010622  \n",
              "2018-08-21                  1                 1         1  0.021021  \n",
              "2018-08-22                  1                 1         1  0.024510  \n",
              "2018-08-23                  1                 1         1  0.066507  \n",
              "2018-08-24                  0                 0         0  0.075819  "
            ],
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>Adj Close</th>\n",
              "      <th>Close</th>\n",
              "      <th>High</th>\n",
              "      <th>Low</th>\n",
              "      <th>Open</th>\n",
              "      <th>Volume</th>\n",
              "      <th>Increase_Decrease</th>\n",
              "      <th>Buy_Sell_on_Open</th>\n",
              "      <th>Buy_Sell</th>\n",
              "      <th>Returns</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>Date</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2018-08-20</th>\n",
              "      <td>19.980000</td>\n",
              "      <td>19.980000</td>\n",
              "      <td>20.08</td>\n",
              "      <td>19.350000</td>\n",
              "      <td>19.790001</td>\n",
              "      <td>62983200</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.010622</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-08-21</th>\n",
              "      <td>20.400000</td>\n",
              "      <td>20.400000</td>\n",
              "      <td>20.42</td>\n",
              "      <td>19.860001</td>\n",
              "      <td>19.980000</td>\n",
              "      <td>55629000</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.021021</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-08-22</th>\n",
              "      <td>20.900000</td>\n",
              "      <td>20.900000</td>\n",
              "      <td>20.92</td>\n",
              "      <td>20.209999</td>\n",
              "      <td>20.280001</td>\n",
              "      <td>62002700</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.024510</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-08-23</th>\n",
              "      <td>22.290001</td>\n",
              "      <td>22.290001</td>\n",
              "      <td>22.32</td>\n",
              "      <td>21.139999</td>\n",
              "      <td>21.190001</td>\n",
              "      <td>113444100</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.066507</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2018-08-24</th>\n",
              "      <td>23.980000</td>\n",
              "      <td>23.980000</td>\n",
              "      <td>24.00</td>\n",
              "      <td>22.670000</td>\n",
              "      <td>22.910000</td>\n",
              "      <td>164328200</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.075819</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ]
          },
          "metadata": {}
        }
      ],
      "execution_count": 4,
      "metadata": {
        "collapsed": true,
        "outputExpanded": false,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:46.703Z",
          "iopub.execute_input": "2020-05-24T05:05:46.710Z",
          "iopub.status.idle": "2020-05-24T05:05:46.725Z",
          "shell.execute_reply": "2020-05-24T05:05:47.298Z"
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X = dataset.iloc[ : , 6:8].values\n",
        "y = dataset.iloc[ : , 8].values"
      ],
      "outputs": [],
      "execution_count": 5,
      "metadata": {
        "collapsed": true,
        "outputExpanded": false,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:46.733Z",
          "iopub.execute_input": "2020-05-24T05:05:46.739Z",
          "iopub.status.idle": "2020-05-24T05:05:46.750Z",
          "shell.execute_reply": "2020-05-24T05:05:47.301Z"
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import train_test_split  \n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20)  "
      ],
      "outputs": [],
      "execution_count": 6,
      "metadata": {
        "collapsed": true,
        "outputExpanded": false,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:46.757Z",
          "iopub.execute_input": "2020-05-24T05:05:46.762Z",
          "iopub.status.idle": "2020-05-24T05:05:47.258Z",
          "shell.execute_reply": "2020-05-24T05:05:47.303Z"
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(X_train.shape)\n",
        "print(y_train.shape)\n",
        "print(X_test.shape)\n",
        "print(y_test.shape)"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(936, 2)\n",
            "(936,)\n",
            "(234, 2)\n",
            "(234,)\n"
          ]
        }
      ],
      "execution_count": 7,
      "metadata": {
        "collapsed": true,
        "outputExpanded": false,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:47.265Z",
          "iopub.execute_input": "2020-05-24T05:05:47.269Z",
          "iopub.status.idle": "2020-05-24T05:05:47.278Z",
          "shell.execute_reply": "2020-05-24T05:05:47.305Z"
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import statsmodels.api as sm\n",
        "from statsmodels.formula.api import poisson\n",
        "from statsmodels.formula.api import negativebinomial"
      ],
      "outputs": [],
      "execution_count": 8,
      "metadata": {
        "collapsed": true,
        "outputExpanded": false,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:47.284Z",
          "iopub.execute_input": "2020-05-24T05:05:47.287Z",
          "iopub.status.idle": "2020-05-24T05:05:47.828Z",
          "shell.execute_reply": "2020-05-24T05:05:47.838Z"
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "m1 = poisson('Increase_Decrease ~ Returns', data =dataset).fit()\n",
        "print(m1.summary())"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Optimization terminated successfully.\n",
            "         Current function value: 0.812951\n",
            "         Iterations 4\n",
            "                          Poisson Regression Results                          \n",
            "==============================================================================\n",
            "Dep. Variable:      Increase_Decrease   No. Observations:                 1170\n",
            "Model:                        Poisson   Df Residuals:                     1168\n",
            "Method:                           MLE   Df Model:                            1\n",
            "Date:                Sat, 23 May 2020   Pseudo R-squ.:               0.0001306\n",
            "Time:                        22:05:47   Log-Likelihood:                -951.15\n",
            "converged:                       True   LL-Null:                       -951.28\n",
            "                                        LLR p-value:                    0.6181\n",
            "==============================================================================\n",
            "                 coef    std err          z      P>|z|      [0.025      0.975]\n",
            "------------------------------------------------------------------------------\n",
            "Intercept     -0.7896      0.043    -18.156      0.000      -0.875      -0.704\n",
            "Returns        0.5577      1.108      0.503      0.615      -1.615       2.730\n",
            "==============================================================================\n"
          ]
        }
      ],
      "execution_count": 9,
      "metadata": {
        "collapsed": true,
        "outputExpanded": false,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:47.847Z",
          "iopub.execute_input": "2020-05-24T05:05:47.851Z",
          "iopub.status.idle": "2020-05-24T05:05:47.868Z",
          "shell.execute_reply": "2020-05-24T05:05:48.197Z"
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model_fit1 = dataset\n",
        "preds_1 = m1.predict()\n",
        "model_fit1['preds'] = preds_1\n",
        "\n",
        "print(model_fit1.head(5))"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "            Adj Close  Close  High   Low  Open    Volume  Increase_Decrease  \\\n",
            "Date                                                                          \n",
            "2014-01-03       4.00   4.00  4.00  3.88  3.98  22887200                  1   \n",
            "2014-01-06       4.13   4.13  4.18  3.99  4.01  42398300                  1   \n",
            "2014-01-07       4.18   4.18  4.25  4.11  4.19  42932100                  0   \n",
            "2014-01-08       4.18   4.18  4.26  4.14  4.23  30678700                  0   \n",
            "2014-01-09       4.09   4.09  4.23  4.05  4.20  30667600                  0   \n",
            "\n",
            "            Buy_Sell_on_Open  Buy_Sell   Returns     preds  \n",
            "Date                                                        \n",
            "2014-01-03                 1         1  0.012658  0.457245  \n",
            "2014-01-06                 1         1  0.032500  0.462334  \n",
            "2014-01-07                 1         0  0.012106  0.457105  \n",
            "2014-01-08                 0         0  0.000000  0.454029  \n",
            "2014-01-09                 0         1 -0.021531  0.448609  \n"
          ]
        }
      ],
      "execution_count": 10,
      "metadata": {
        "collapsed": true,
        "outputExpanded": false,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:47.877Z",
          "iopub.execute_input": "2020-05-24T05:05:47.880Z",
          "iopub.status.idle": "2020-05-24T05:05:47.938Z",
          "shell.execute_reply": "2020-05-24T05:05:48.200Z"
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "columns = ['Adj Close', 'High', 'Low', 'Open']\n",
        "X = [s for s in dataset[columns].values]\n",
        "X = sm.add_constant(X, prepend = False)\n",
        "Y = [s for s in dataset['Returns'].values]"
      ],
      "outputs": [],
      "execution_count": 11,
      "metadata": {
        "collapsed": true,
        "outputExpanded": false,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:47.947Z",
          "iopub.execute_input": "2020-05-24T05:05:47.951Z",
          "iopub.status.idle": "2020-05-24T05:05:47.959Z",
          "shell.execute_reply": "2020-05-24T05:05:48.202Z"
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "m2 = sm.Poisson(Y, X).fit()\n",
        "print(m2.summary())"
      ],
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Warning: Maximum number of iterations has been exceeded.\n",
            "         Current function value: 6812363.638260\n",
            "         Iterations: 35\n",
            "                          Poisson Regression Results                          \n",
            "==============================================================================\n",
            "Dep. Variable:                      y   No. Observations:                 1170\n",
            "Model:                        Poisson   Df Residuals:                     1165\n",
            "Method:                           MLE   Df Model:                            4\n",
            "Date:                Sat, 23 May 2020   Pseudo R-squ.:              -4.317e+08\n",
            "Time:                        22:05:47   Log-Likelihood:            -7.9705e+09\n",
            "converged:                      False   LL-Null:                       -18.462\n",
            "                                        LLR p-value:                     1.000\n",
            "==============================================================================\n",
            "                 coef    std err          z      P>|z|      [0.025      0.975]\n",
            "------------------------------------------------------------------------------\n",
            "x1            46.6108      0.018   2624.528      0.000      46.576      46.646\n",
            "x2             7.1890      0.013    540.065      0.000       7.163       7.215\n",
            "x3            -7.7512      0.003  -2643.728      0.000      -7.757      -7.745\n",
            "x4           -46.2801      0.004  -1.11e+04      0.000     -46.288     -46.272\n",
            "const        -40.7364      0.005  -7809.609      0.000     -40.747     -40.726\n",
            "==============================================================================\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "C:\\Users\\Tin Hang\\Anaconda3\\envs\\py35\\lib\\site-packages\\statsmodels\\base\\model.py:508: ConvergenceWarning: Maximum Likelihood optimization failed to converge. Check mle_retvals\n",
            "  \"Check mle_retvals\", ConvergenceWarning)\n"
          ]
        }
      ],
      "execution_count": 12,
      "metadata": {
        "collapsed": true,
        "outputExpanded": false,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:47.966Z",
          "iopub.execute_input": "2020-05-24T05:05:47.969Z",
          "iopub.status.idle": "2020-05-24T05:05:48.004Z",
          "shell.execute_reply": "2020-05-24T05:05:48.204Z"
        }
      }
    },
    {
      "cell_type": "code",
      "source": [
        "preds = m1.predict()\n",
        "plt.plot(range(len(Y)), Y, 'r*-', range(len(Y)), preds, 'bo-')\n",
        "plt.title('True values vs fitted values')\n",
        "plt.legend(['Real Values', 'Fitted Values'])\n",
        "plt.show()"
      ],
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": [
              "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\n"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "execution_count": 13,
      "metadata": {
        "collapsed": true,
        "outputExpanded": false,
        "jupyter": {
          "source_hidden": false,
          "outputs_hidden": false
        },
        "nteract": {
          "transient": {
            "deleting": false
          }
        },
        "execution": {
          "iopub.status.busy": "2020-05-24T05:05:48.011Z",
          "iopub.execute_input": "2020-05-24T05:05:48.028Z",
          "iopub.status.idle": "2020-05-24T05:05:48.188Z",
          "shell.execute_reply": "2020-05-24T05:05:48.206Z"
        }
      }
    }
  ],
  "metadata": {
    "kernel_info": {
      "name": "python3"
    },
    "language_info": {
      "name": "python",
      "file_extension": ".py",
      "pygments_lexer": "ipython3",
      "nbconvert_exporter": "python",
      "mimetype": "text/x-python",
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "version": "3.5.5"
    },
    "kernelspec": {
      "name": "python3",
      "language": "python",
      "display_name": "Python 3"
    },
    "nteract": {
      "version": "0.23.1"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 4
}