pytorch-binary-classification.ipynb

{"metadata":{"kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"language_info":{"name":"python","version":"3.10.10","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat_minor":5,"nbformat":4,"cells":[{"source":"<a href=\"https://www.kaggle.com/code/averma111/pytorch-binary-classification?scriptVersionId=128473923\" target=\"_blank\"><img align=\"left\" alt=\"Kaggle\" title=\"Open in Kaggle\" src=\"https://kaggle.com/static/images/open-in-kaggle.svg\"></a>","metadata":{},"cell_type":"markdown"},{"cell_type":"code","source":"# importing libraries\n\nimport numpy as np\nimport pandas as pd\nimport seaborn as sns\nimport matplotlib.pyplot as plt\nimport warnings\nwarnings.filterwarnings('ignore')","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.216014Z","iopub.execute_input":"2023-05-06T00:57:53.216385Z","iopub.status.idle":"2023-05-06T00:57:53.222143Z","shell.execute_reply.started":"2023-05-06T00:57:53.216355Z","shell.execute_reply":"2023-05-06T00:57:53.220727Z"},"trusted":true},"execution_count":27,"outputs":[]},{"cell_type":"code","source":"# Reading the dataset\n\ntrain_spine = pd.read_csv(\"/kaggle/input/lower-back-pain-symptoms-datasetlabelled/Dataset_spine.csv\")\ntrain_spine.head()","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.229943Z","iopub.execute_input":"2023-05-06T00:57:53.230444Z","iopub.status.idle":"2023-05-06T00:57:53.259438Z","shell.execute_reply.started":"2023-05-06T00:57:53.230399Z","shell.execute_reply":"2023-05-06T00:57:53.258137Z"},"trusted":true},"execution_count":28,"outputs":[{"execution_count":28,"output_type":"execute_result","data":{"text/plain":"   pelvic_incidence  pelvic_tilt  lumbar_lordosis_angle  sacral_slope  \\\n0         63.027817    22.552586              39.609117     40.475232   \n1         39.056951    10.060991              25.015378     28.995960   \n2         68.832021    22.218482              50.092194     46.613539   \n3         69.297008    24.652878              44.311238     44.644130   \n4         49.712859     9.652075              28.317406     40.060784   \n\n   pelvic_radius  degree_spondylolisthesis  pelvic_slope  direct_tilt  \\\n0      98.672917                 -0.254400      0.744503      12.5661   \n1     114.405425                  4.564259      0.415186      12.8874   \n2     105.985135                 -3.530317      0.474889      26.8343   \n3     101.868495                 11.211523      0.369345      23.5603   \n4     108.168725                  7.918501      0.543360      35.4940   \n\n   thoracic_slope  cervical_tilt  sacrum_angle  scoliosis_slope Class_att  \n0         14.5386       15.30468    -28.658501          43.5123  Abnormal  \n1         17.5323       16.78486    -25.530607          16.1102  Abnormal  \n2         17.4861       16.65897    -29.031888          19.2221  Abnormal  \n3         12.7074       11.42447    -30.470246          18.8329  Abnormal  \n4         15.9546        8.87237    -16.378376          24.9171  Abnormal  ","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>pelvic_incidence</th>\n      <th>pelvic_tilt</th>\n      <th>lumbar_lordosis_angle</th>\n      <th>sacral_slope</th>\n      <th>pelvic_radius</th>\n      <th>degree_spondylolisthesis</th>\n      <th>pelvic_slope</th>\n      <th>direct_tilt</th>\n      <th>thoracic_slope</th>\n      <th>cervical_tilt</th>\n      <th>sacrum_angle</th>\n      <th>scoliosis_slope</th>\n      <th>Class_att</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>63.027817</td>\n      <td>22.552586</td>\n      <td>39.609117</td>\n      <td>40.475232</td>\n      <td>98.672917</td>\n      <td>-0.254400</td>\n      <td>0.744503</td>\n      <td>12.5661</td>\n      <td>14.5386</td>\n      <td>15.30468</td>\n      <td>-28.658501</td>\n      <td>43.5123</td>\n      <td>Abnormal</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>39.056951</td>\n      <td>10.060991</td>\n      <td>25.015378</td>\n      <td>28.995960</td>\n      <td>114.405425</td>\n      <td>4.564259</td>\n      <td>0.415186</td>\n      <td>12.8874</td>\n      <td>17.5323</td>\n      <td>16.78486</td>\n      <td>-25.530607</td>\n      <td>16.1102</td>\n      <td>Abnormal</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>68.832021</td>\n      <td>22.218482</td>\n      <td>50.092194</td>\n      <td>46.613539</td>\n      <td>105.985135</td>\n      <td>-3.530317</td>\n      <td>0.474889</td>\n      <td>26.8343</td>\n      <td>17.4861</td>\n      <td>16.65897</td>\n      <td>-29.031888</td>\n      <td>19.2221</td>\n      <td>Abnormal</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>69.297008</td>\n      <td>24.652878</td>\n      <td>44.311238</td>\n      <td>44.644130</td>\n      <td>101.868495</td>\n      <td>11.211523</td>\n      <td>0.369345</td>\n      <td>23.5603</td>\n      <td>12.7074</td>\n      <td>11.42447</td>\n      <td>-30.470246</td>\n      <td>18.8329</td>\n      <td>Abnormal</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>49.712859</td>\n      <td>9.652075</td>\n      <td>28.317406</td>\n      <td>40.060784</td>\n      <td>108.168725</td>\n      <td>7.918501</td>\n      <td>0.543360</td>\n      <td>35.4940</td>\n      <td>15.9546</td>\n      <td>8.87237</td>\n      <td>-16.378376</td>\n      <td>24.9171</td>\n      <td>Abnormal</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"cell_type":"code","source":"train_spine.info()","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.262223Z","iopub.execute_input":"2023-05-06T00:57:53.262663Z","iopub.status.idle":"2023-05-06T00:57:53.276477Z","shell.execute_reply.started":"2023-05-06T00:57:53.262621Z","shell.execute_reply":"2023-05-06T00:57:53.275391Z"},"trusted":true},"execution_count":29,"outputs":[{"name":"stdout","text":"<class 'pandas.core.frame.DataFrame'>\nRangeIndex: 310 entries, 0 to 309\nData columns (total 13 columns):\n #   Column                    Non-Null Count  Dtype  \n---  ------                    --------------  -----  \n 0   pelvic_incidence          310 non-null    float64\n 1   pelvic_tilt               310 non-null    float64\n 2   lumbar_lordosis_angle     310 non-null    float64\n 3   sacral_slope              310 non-null    float64\n 4   pelvic_radius             310 non-null    float64\n 5   degree_spondylolisthesis  310 non-null    float64\n 6   pelvic_slope              310 non-null    float64\n 7   direct_tilt               310 non-null    float64\n 8   thoracic_slope            310 non-null    float64\n 9   cervical_tilt             310 non-null    float64\n 10  sacrum_angle              310 non-null    float64\n 11  scoliosis_slope           310 non-null    float64\n 12  Class_att                 310 non-null    object \ndtypes: float64(12), object(1)\nmemory usage: 31.6+ KB\n","output_type":"stream"}]},{"cell_type":"code","source":"def summary(text, df):\n    print(f'{text} shape: {df.shape}')\n    summ = pd.DataFrame(df.dtypes, columns=['dtypes'])\n    summ['null'] = df.isnull().sum()\n    summ['unique'] = df.nunique()\n    summ['min'] = df.min()\n    summ['median'] = df.median()\n    summ['max'] = df.max()\n    summ['mean'] = df.mean()\n    summ['std'] = df.std()\n    return summ","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.277819Z","iopub.execute_input":"2023-05-06T00:57:53.278897Z","iopub.status.idle":"2023-05-06T00:57:53.288668Z","shell.execute_reply.started":"2023-05-06T00:57:53.278852Z","shell.execute_reply":"2023-05-06T00:57:53.287282Z"},"trusted":true},"execution_count":30,"outputs":[]},{"cell_type":"code","source":"summary('train_spine',train_spine)","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.290127Z","iopub.execute_input":"2023-05-06T00:57:53.290898Z","iopub.status.idle":"2023-05-06T00:57:53.32169Z","shell.execute_reply.started":"2023-05-06T00:57:53.290864Z","shell.execute_reply":"2023-05-06T00:57:53.320918Z"},"trusted":true},"execution_count":31,"outputs":[{"name":"stdout","text":"train_spine shape: (310, 13)\n","output_type":"stream"},{"execution_count":31,"output_type":"execute_result","data":{"text/plain":"                           dtypes  null  unique        min      median  \\\npelvic_incidence          float64     0     310  26.147921   58.691038   \npelvic_tilt               float64     0     310  -6.554948   16.357689   \nlumbar_lordosis_angle     float64     0     280       14.0   49.562398   \nsacral_slope              float64     0     281  13.366931   42.404912   \npelvic_radius             float64     0     310  70.082575  118.268178   \ndegree_spondylolisthesis  float64     0     310 -11.058179   11.767934   \npelvic_slope              float64     0     310    0.00322    0.475989   \ndirect_tilt               float64     0     310      7.027   21.907150   \nthoracic_slope            float64     0     310     7.0378   12.938450   \ncervical_tilt             float64     0     310     7.0306   11.953835   \nsacrum_angle              float64     0     310 -35.287375  -14.622856   \nscoliosis_slope           float64     0     310     7.0079   24.931950   \nClass_att                  object     0       2   Abnormal         NaN   \n\n                                 max        mean        std  \npelvic_incidence          129.834041   60.496653  17.236520  \npelvic_tilt                49.431864   17.542822  10.008330  \nlumbar_lordosis_angle     125.742385   51.930930  18.554064  \nsacral_slope              121.429566   42.953831  13.423102  \npelvic_radius             163.071041  117.920655  13.317377  \ndegree_spondylolisthesis  418.543082   26.296694  37.559027  \npelvic_slope                0.998827    0.472979   0.285787  \ndirect_tilt                  36.7439   21.321526   8.639423  \nthoracic_slope                19.324   13.064511   3.399713  \ncervical_tilt               16.82108   11.933317   2.893265  \nsacrum_angle                6.972071  -14.053139  12.225582  \nscoliosis_slope              44.3412   25.645981  10.450558  \nClass_att                     Normal         NaN        NaN  ","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>dtypes</th>\n      <th>null</th>\n      <th>unique</th>\n      <th>min</th>\n      <th>median</th>\n      <th>max</th>\n      <th>mean</th>\n      <th>std</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>pelvic_incidence</th>\n      <td>float64</td>\n      <td>0</td>\n      <td>310</td>\n      <td>26.147921</td>\n      <td>58.691038</td>\n      <td>129.834041</td>\n      <td>60.496653</td>\n      <td>17.236520</td>\n    </tr>\n    <tr>\n      <th>pelvic_tilt</th>\n      <td>float64</td>\n      <td>0</td>\n      <td>310</td>\n      <td>-6.554948</td>\n      <td>16.357689</td>\n      <td>49.431864</td>\n      <td>17.542822</td>\n      <td>10.008330</td>\n    </tr>\n    <tr>\n      <th>lumbar_lordosis_angle</th>\n      <td>float64</td>\n      <td>0</td>\n      <td>280</td>\n      <td>14.0</td>\n      <td>49.562398</td>\n      <td>125.742385</td>\n      <td>51.930930</td>\n      <td>18.554064</td>\n    </tr>\n    <tr>\n      <th>sacral_slope</th>\n      <td>float64</td>\n      <td>0</td>\n      <td>281</td>\n      <td>13.366931</td>\n      <td>42.404912</td>\n      <td>121.429566</td>\n      <td>42.953831</td>\n      <td>13.423102</td>\n    </tr>\n    <tr>\n      <th>pelvic_radius</th>\n      <td>float64</td>\n      <td>0</td>\n      <td>310</td>\n      <td>70.082575</td>\n      <td>118.268178</td>\n      <td>163.071041</td>\n      <td>117.920655</td>\n      <td>13.317377</td>\n    </tr>\n    <tr>\n      <th>degree_spondylolisthesis</th>\n      <td>float64</td>\n      <td>0</td>\n      <td>310</td>\n      <td>-11.058179</td>\n      <td>11.767934</td>\n      <td>418.543082</td>\n      <td>26.296694</td>\n      <td>37.559027</td>\n    </tr>\n    <tr>\n      <th>pelvic_slope</th>\n      <td>float64</td>\n      <td>0</td>\n      <td>310</td>\n      <td>0.00322</td>\n      <td>0.475989</td>\n      <td>0.998827</td>\n      <td>0.472979</td>\n      <td>0.285787</td>\n    </tr>\n    <tr>\n      <th>direct_tilt</th>\n      <td>float64</td>\n      <td>0</td>\n      <td>310</td>\n      <td>7.027</td>\n      <td>21.907150</td>\n      <td>36.7439</td>\n      <td>21.321526</td>\n      <td>8.639423</td>\n    </tr>\n    <tr>\n      <th>thoracic_slope</th>\n      <td>float64</td>\n      <td>0</td>\n      <td>310</td>\n      <td>7.0378</td>\n      <td>12.938450</td>\n      <td>19.324</td>\n      <td>13.064511</td>\n      <td>3.399713</td>\n    </tr>\n    <tr>\n      <th>cervical_tilt</th>\n      <td>float64</td>\n      <td>0</td>\n      <td>310</td>\n      <td>7.0306</td>\n      <td>11.953835</td>\n      <td>16.82108</td>\n      <td>11.933317</td>\n      <td>2.893265</td>\n    </tr>\n    <tr>\n      <th>sacrum_angle</th>\n      <td>float64</td>\n      <td>0</td>\n      <td>310</td>\n      <td>-35.287375</td>\n      <td>-14.622856</td>\n      <td>6.972071</td>\n      <td>-14.053139</td>\n      <td>12.225582</td>\n    </tr>\n    <tr>\n      <th>scoliosis_slope</th>\n      <td>float64</td>\n      <td>0</td>\n      <td>310</td>\n      <td>7.0079</td>\n      <td>24.931950</td>\n      <td>44.3412</td>\n      <td>25.645981</td>\n      <td>10.450558</td>\n    </tr>\n    <tr>\n      <th>Class_att</th>\n      <td>object</td>\n      <td>0</td>\n      <td>2</td>\n      <td>Abnormal</td>\n      <td>NaN</td>\n      <td>Normal</td>\n      <td>NaN</td>\n      <td>NaN</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"cell_type":"code","source":"# Check number of Normal and Abnormal entires\nsns.countplot(x=train_spine['Class_att'], data=train_spine)","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.324045Z","iopub.execute_input":"2023-05-06T00:57:53.324609Z","iopub.status.idle":"2023-05-06T00:57:53.535785Z","shell.execute_reply.started":"2023-05-06T00:57:53.324579Z","shell.execute_reply":"2023-05-06T00:57:53.534858Z"},"trusted":true},"execution_count":32,"outputs":[{"execution_count":32,"output_type":"execute_result","data":{"text/plain":"<AxesSubplot: xlabel='Class_att', ylabel='count'>"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"<Figure size 640x480 with 1 Axes>","image/png":"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"},"metadata":{}}]},{"cell_type":"code","source":"# Encode the target column \n# PyTorch supports labels starting from 0. That is [0, n]. We need to remap our labels to start from 0.\n\nencoded_target = {\n    'Abnormal':0,\n    'Normal':1\n}\n\n# Cast target column as category\ntrain_spine['Class_att'] = train_spine['Class_att'].astype('category')\n\n#Replace the Abnorml and Normal with encoded_target\ntrain_spine['Class_att'].replace(encoded_target,inplace=True)\n","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.537379Z","iopub.execute_input":"2023-05-06T00:57:53.537743Z","iopub.status.idle":"2023-05-06T00:57:53.54563Z","shell.execute_reply.started":"2023-05-06T00:57:53.537701Z","shell.execute_reply":"2023-05-06T00:57:53.544326Z"},"trusted":true},"execution_count":33,"outputs":[]},{"cell_type":"code","source":"train_spine.head()","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.546653Z","iopub.execute_input":"2023-05-06T00:57:53.547001Z","iopub.status.idle":"2023-05-06T00:57:53.574086Z","shell.execute_reply.started":"2023-05-06T00:57:53.546946Z","shell.execute_reply":"2023-05-06T00:57:53.572851Z"},"trusted":true},"execution_count":34,"outputs":[{"execution_count":34,"output_type":"execute_result","data":{"text/plain":"   pelvic_incidence  pelvic_tilt  lumbar_lordosis_angle  sacral_slope  \\\n0         63.027817    22.552586              39.609117     40.475232   \n1         39.056951    10.060991              25.015378     28.995960   \n2         68.832021    22.218482              50.092194     46.613539   \n3         69.297008    24.652878              44.311238     44.644130   \n4         49.712859     9.652075              28.317406     40.060784   \n\n   pelvic_radius  degree_spondylolisthesis  pelvic_slope  direct_tilt  \\\n0      98.672917                 -0.254400      0.744503      12.5661   \n1     114.405425                  4.564259      0.415186      12.8874   \n2     105.985135                 -3.530317      0.474889      26.8343   \n3     101.868495                 11.211523      0.369345      23.5603   \n4     108.168725                  7.918501      0.543360      35.4940   \n\n   thoracic_slope  cervical_tilt  sacrum_angle  scoliosis_slope Class_att  \n0         14.5386       15.30468    -28.658501          43.5123         0  \n1         17.5323       16.78486    -25.530607          16.1102         0  \n2         17.4861       16.65897    -29.031888          19.2221         0  \n3         12.7074       11.42447    -30.470246          18.8329         0  \n4         15.9546        8.87237    -16.378376          24.9171         0  ","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>pelvic_incidence</th>\n      <th>pelvic_tilt</th>\n      <th>lumbar_lordosis_angle</th>\n      <th>sacral_slope</th>\n      <th>pelvic_radius</th>\n      <th>degree_spondylolisthesis</th>\n      <th>pelvic_slope</th>\n      <th>direct_tilt</th>\n      <th>thoracic_slope</th>\n      <th>cervical_tilt</th>\n      <th>sacrum_angle</th>\n      <th>scoliosis_slope</th>\n      <th>Class_att</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>63.027817</td>\n      <td>22.552586</td>\n      <td>39.609117</td>\n      <td>40.475232</td>\n      <td>98.672917</td>\n      <td>-0.254400</td>\n      <td>0.744503</td>\n      <td>12.5661</td>\n      <td>14.5386</td>\n      <td>15.30468</td>\n      <td>-28.658501</td>\n      <td>43.5123</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>39.056951</td>\n      <td>10.060991</td>\n      <td>25.015378</td>\n      <td>28.995960</td>\n      <td>114.405425</td>\n      <td>4.564259</td>\n      <td>0.415186</td>\n      <td>12.8874</td>\n      <td>17.5323</td>\n      <td>16.78486</td>\n      <td>-25.530607</td>\n      <td>16.1102</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>68.832021</td>\n      <td>22.218482</td>\n      <td>50.092194</td>\n      <td>46.613539</td>\n      <td>105.985135</td>\n      <td>-3.530317</td>\n      <td>0.474889</td>\n      <td>26.8343</td>\n      <td>17.4861</td>\n      <td>16.65897</td>\n      <td>-29.031888</td>\n      <td>19.2221</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>69.297008</td>\n      <td>24.652878</td>\n      <td>44.311238</td>\n      <td>44.644130</td>\n      <td>101.868495</td>\n      <td>11.211523</td>\n      <td>0.369345</td>\n      <td>23.5603</td>\n      <td>12.7074</td>\n      <td>11.42447</td>\n      <td>-30.470246</td>\n      <td>18.8329</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>49.712859</td>\n      <td>9.652075</td>\n      <td>28.317406</td>\n      <td>40.060784</td>\n      <td>108.168725</td>\n      <td>7.918501</td>\n      <td>0.543360</td>\n      <td>35.4940</td>\n      <td>15.9546</td>\n      <td>8.87237</td>\n      <td>-16.378376</td>\n      <td>24.9171</td>\n      <td>0</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"cell_type":"code","source":"# Define the features and target column\nX = train_spine.iloc[:,0:-1]\ny = train_spine.iloc[:,-1]","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.575786Z","iopub.execute_input":"2023-05-06T00:57:53.576597Z","iopub.status.idle":"2023-05-06T00:57:53.584323Z","shell.execute_reply.started":"2023-05-06T00:57:53.576563Z","shell.execute_reply":"2023-05-06T00:57:53.583071Z"},"trusted":true},"execution_count":35,"outputs":[]},{"cell_type":"code","source":"# Split the data into training and test \nfrom sklearn.model_selection import train_test_split\nX_train,X_test,y_train,y_test = train_test_split(X,y,test_size=0.2,random_state=42)\nX_train.shape,X_test.shape,y_train.shape,y_test.shape","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.585905Z","iopub.execute_input":"2023-05-06T00:57:53.586426Z","iopub.status.idle":"2023-05-06T00:57:53.604202Z","shell.execute_reply.started":"2023-05-06T00:57:53.586381Z","shell.execute_reply":"2023-05-06T00:57:53.602949Z"},"trusted":true},"execution_count":36,"outputs":[{"execution_count":36,"output_type":"execute_result","data":{"text/plain":"((248, 12), (62, 12), (248,), (62,))"},"metadata":{}}]},{"cell_type":"code","source":"## Standardize/Normalize the inputs =>https://www.jeremyjordan.me/batch-normalization/\n#For neural networks to train properly, we need to standardize the input values. \n#We standardize features by removing the mean and scaling to unit variance. T\n#he standard score of a sample x where the mean is u and the standard deviation is s is calculated as\n#Z = (X - u) / S\nfrom sklearn.preprocessing import StandardScaler   \nscaler = StandardScaler()\nX_train = scaler.fit_transform(X_train)\nX_test = scaler.fit_transform(X_test)","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.606334Z","iopub.execute_input":"2023-05-06T00:57:53.610514Z","iopub.status.idle":"2023-05-06T00:57:53.62469Z","shell.execute_reply.started":"2023-05-06T00:57:53.610476Z","shell.execute_reply":"2023-05-06T00:57:53.623228Z"},"trusted":true},"execution_count":37,"outputs":[]},{"cell_type":"code","source":"# Model Parameters \nEPOCHS = 500\nBATCH_SIZE = 32\nLEARNING_RATE = 1e-4","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.626217Z","iopub.execute_input":"2023-05-06T00:57:53.626889Z","iopub.status.idle":"2023-05-06T00:57:53.635463Z","shell.execute_reply.started":"2023-05-06T00:57:53.626848Z","shell.execute_reply":"2023-05-06T00:57:53.634221Z"},"trusted":true},"execution_count":38,"outputs":[]},{"cell_type":"code","source":"# Define the DataLoader for train and test data\nfrom torch.utils.data import Dataset, DataLoader\n\n\n# Train Data\nclass TrainData(Dataset):\n    def __init__(self, X_data, y_data):\n        self.X_data = X_data\n        self.y_data = y_data\n        \n    def __getitem__(self, index):\n        return self.X_data[index], self.y_data[index]\n        \n    def __len__ (self):\n        return len(self.X_data)\n    \n# Test Data   \nclass TestData(Dataset):\n    \n    def __init__(self, X_data):\n        self.X_data = X_data\n        \n    def __getitem__(self, index):\n        return self.X_data[index]\n        \n    def __len__ (self):\n        return len(self.X_data)\n    \n    \n","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.637313Z","iopub.execute_input":"2023-05-06T00:57:53.638182Z","iopub.status.idle":"2023-05-06T00:57:53.648369Z","shell.execute_reply.started":"2023-05-06T00:57:53.638136Z","shell.execute_reply":"2023-05-06T00:57:53.647065Z"},"trusted":true},"execution_count":39,"outputs":[]},{"cell_type":"code","source":"# Instantiate the Train and Test data class\nimport torch\ntrain_data = TrainData(torch.FloatTensor(X_train), \n                       torch.FloatTensor(y_train))\ntest_data = TestData(torch.FloatTensor(X_test))","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.650465Z","iopub.execute_input":"2023-05-06T00:57:53.65131Z","iopub.status.idle":"2023-05-06T00:57:53.66071Z","shell.execute_reply.started":"2023-05-06T00:57:53.651264Z","shell.execute_reply":"2023-05-06T00:57:53.659798Z"},"trusted":true},"execution_count":40,"outputs":[]},{"cell_type":"code","source":"# Initialize the DataLoader \ntrain_loader = DataLoader(dataset=train_data, batch_size=BATCH_SIZE, shuffle=True)\ntest_loader = DataLoader(dataset=test_data, batch_size=1)","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.666507Z","iopub.execute_input":"2023-05-06T00:57:53.666887Z","iopub.status.idle":"2023-05-06T00:57:53.67326Z","shell.execute_reply.started":"2023-05-06T00:57:53.666853Z","shell.execute_reply":"2023-05-06T00:57:53.671805Z"},"trusted":true},"execution_count":41,"outputs":[]},{"cell_type":"markdown","source":"### The only thing you need to ensure is that number of output features of one layer should be equal to the input features of the next layer.","metadata":{}},{"cell_type":"markdown","source":"#### The Dropout layer randomly sets input units to 0 with a frequency of rate at each step during training time, which helps prevent overfitting. Inputs not set to 0 are scaled up by 1/(1 - rate) such that the sum over all inputs is unchanged.","metadata":{}},{"cell_type":"markdown","source":"#### Batch Norm is just another network layer that gets inserted between a hidden layer and the next hidden layer. Its job is to take the outputs from the first hidden layer and normalize them before passing them on as the input of the next hidden layer.","metadata":{}},{"cell_type":"code","source":"import torch\n\nclass BinaryClassificationNNet(torch.nn.Module):\n    def __init__(self,input_features):\n        super(BinaryClassificationNNet,self).__init__()\n        # Number of input features is 12.\n        self.input_layer = torch.nn.Linear(input_features,64)\n        self.dense_layer = torch.nn.Linear(64,64)\n        self.output_layer = torch.nn.Linear(64,1)\n        \n        self.relu = torch.nn.ReLU()\n        self.dropout = torch.nn.Dropout(p=0.1)\n        self.batchnorm_1 = torch.nn.BatchNorm1d(64)\n        self.batchnorm_2 = torch.nn.BatchNorm1d(64)\n        \n    \n    def forward(self,inputs):\n        x = self.relu(self.input_layer(inputs))\n        x = self.batchnorm_1(x)\n        x = self.relu(self.dense_layer(x))\n        x = self.batchnorm_2(x)\n        x = self.dropout(x)\n        x = self.output_layer(x)\n        \n        return x","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.675303Z","iopub.execute_input":"2023-05-06T00:57:53.675741Z","iopub.status.idle":"2023-05-06T00:57:53.68726Z","shell.execute_reply.started":"2023-05-06T00:57:53.675697Z","shell.execute_reply":"2023-05-06T00:57:53.686047Z"},"trusted":true},"execution_count":42,"outputs":[]},{"cell_type":"code","source":"device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\nprint(device)","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.68887Z","iopub.execute_input":"2023-05-06T00:57:53.689436Z","iopub.status.idle":"2023-05-06T00:57:53.705434Z","shell.execute_reply.started":"2023-05-06T00:57:53.689389Z","shell.execute_reply":"2023-05-06T00:57:53.704269Z"},"trusted":true},"execution_count":43,"outputs":[{"name":"stdout","text":"cpu\n","output_type":"stream"}]},{"cell_type":"code","source":"model = BinaryClassificationNNet(X_train.shape[1])\nmodel.to(device)\nprint(model)","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.707017Z","iopub.execute_input":"2023-05-06T00:57:53.707815Z","iopub.status.idle":"2023-05-06T00:57:53.719875Z","shell.execute_reply.started":"2023-05-06T00:57:53.70777Z","shell.execute_reply":"2023-05-06T00:57:53.718775Z"},"trusted":true},"execution_count":44,"outputs":[{"name":"stdout","text":"BinaryClassificationNNet(\n  (input_layer): Linear(in_features=12, out_features=64, bias=True)\n  (dense_layer): Linear(in_features=64, out_features=64, bias=True)\n  (output_layer): Linear(in_features=64, out_features=1, bias=True)\n  (relu): ReLU()\n  (dropout): Dropout(p=0.1, inplace=False)\n  (batchnorm_1): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n  (batchnorm_2): BatchNorm1d(64, eps=1e-05, momentum=0.1, affine=True, track_running_stats=True)\n)\n","output_type":"stream"}]},{"cell_type":"code","source":"# Define the loss and optimizer \ncriterion = torch.nn.BCEWithLogitsLoss()\noptimizer = torch.optim.Adam(model.parameters(), lr=LEARNING_RATE)","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.721631Z","iopub.execute_input":"2023-05-06T00:57:53.722861Z","iopub.status.idle":"2023-05-06T00:57:53.731424Z","shell.execute_reply.started":"2023-05-06T00:57:53.722815Z","shell.execute_reply":"2023-05-06T00:57:53.729602Z"},"trusted":true},"execution_count":45,"outputs":[]},{"cell_type":"code","source":"# Define the accuracy function \n\ndef binary_acc(y_pred, y_test):\n    y_pred_tag = torch.round(torch.sigmoid(y_pred))\n    correct_results_sum = (y_pred_tag == y_test).sum().float()\n    accuracy = correct_results_sum/y_test.shape[0]\n    accuracy = torch.round(accuracy * 100)\n    \n    return accuracy","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.733445Z","iopub.execute_input":"2023-05-06T00:57:53.734614Z","iopub.status.idle":"2023-05-06T00:57:53.744302Z","shell.execute_reply.started":"2023-05-06T00:57:53.734569Z","shell.execute_reply":"2023-05-06T00:57:53.742536Z"},"trusted":true},"execution_count":46,"outputs":[]},{"cell_type":"code","source":"# Tain the model\nfrom tqdm.notebook import tqdm\n\nmodel.train()\nfor e in tqdm(range(1, EPOCHS+1)):\n    epoch_loss = 0\n    epoch_acc = 0\n    \n    for X_batch, y_batch in train_loader:\n        X_batch, y_batch = X_batch.to(device), y_batch.to(device)\n        optimizer.zero_grad()\n        \n        y_pred = model(X_batch)\n        #print(y_batch.unsqueeze(1))\n        loss = criterion(y_pred, y_batch.unsqueeze(1))\n        acc = binary_acc(y_pred, y_batch.unsqueeze(1))\n        \n        loss.backward()\n        optimizer.step()\n        \n        epoch_loss += loss.item()\n        epoch_acc += acc.item()\n        \n    if e%100==0:\n        print(f'Epoch {e+0:03}: | Loss: {epoch_loss/len(train_loader):.5f} | Acc: {epoch_acc/len(train_loader):.3f}')","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:57:53.746089Z","iopub.execute_input":"2023-05-06T00:57:53.746933Z","iopub.status.idle":"2023-05-06T00:58:02.785253Z","shell.execute_reply.started":"2023-05-06T00:57:53.746887Z","shell.execute_reply":"2023-05-06T00:58:02.784515Z"},"trusted":true},"execution_count":47,"outputs":[{"output_type":"display_data","data":{"text/plain":"  0%|          | 0/500 [00:00<?, ?it/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"943ce227bb4e40fda48cf9703aff4fc2"}},"metadata":{}},{"name":"stdout","text":"Epoch 100: | Loss: 0.30448 | Acc: 89.375\nEpoch 200: | Loss: 0.14708 | Acc: 95.875\nEpoch 300: | Loss: 0.09285 | Acc: 98.500\nEpoch 400: | Loss: 0.06064 | Acc: 98.875\nEpoch 500: | Loss: 0.03174 | Acc: 99.625\n","output_type":"stream"}]},{"cell_type":"code","source":"# Evaluate the model\n\ny_pred_list = []\nmodel.eval()\nwith torch.no_grad():\n    for X_batch in test_loader:\n        X_batch = X_batch.to(device)\n        y_test_pred = model(X_batch)\n        y_test_pred = torch.sigmoid(y_test_pred)\n        y_pred_tag = torch.round(y_test_pred)\n        y_pred_list.append(y_pred_tag.cpu().numpy())\n\ny_pred_list = [a.squeeze().tolist() for a in y_pred_list]","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:58:02.786406Z","iopub.execute_input":"2023-05-06T00:58:02.787182Z","iopub.status.idle":"2023-05-06T00:58:02.806126Z","shell.execute_reply.started":"2023-05-06T00:58:02.787151Z","shell.execute_reply":"2023-05-06T00:58:02.805258Z"},"trusted":true},"execution_count":48,"outputs":[]},{"cell_type":"markdown","source":"### Confusion Matrix","metadata":{}},{"cell_type":"code","source":"from sklearn.metrics import confusion_matrix, classification_report\nconfusion_matrix(y_test, y_pred_list)","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:58:02.807376Z","iopub.execute_input":"2023-05-06T00:58:02.807687Z","iopub.status.idle":"2023-05-06T00:58:02.820635Z","shell.execute_reply.started":"2023-05-06T00:58:02.807657Z","shell.execute_reply":"2023-05-06T00:58:02.81942Z"},"trusted":true},"execution_count":49,"outputs":[{"execution_count":49,"output_type":"execute_result","data":{"text/plain":"array([[36,  8],\n       [ 3, 15]])"},"metadata":{}}]},{"cell_type":"markdown","source":"### Classification Report","metadata":{}},{"cell_type":"code","source":"print(classification_report(y_test, y_pred_list))","metadata":{"execution":{"iopub.status.busy":"2023-05-06T00:58:02.822308Z","iopub.execute_input":"2023-05-06T00:58:02.822634Z","iopub.status.idle":"2023-05-06T00:58:02.835905Z","shell.execute_reply.started":"2023-05-06T00:58:02.822603Z","shell.execute_reply":"2023-05-06T00:58:02.834863Z"},"trusted":true},"execution_count":50,"outputs":[{"name":"stdout","text":"              precision    recall  f1-score   support\n\n           0       0.92      0.82      0.87        44\n           1       0.65      0.83      0.73        18\n\n    accuracy                           0.82        62\n   macro avg       0.79      0.83      0.80        62\nweighted avg       0.84      0.82      0.83        62\n\n","output_type":"stream"}]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}