{
  "id": 328508,
  "title": "Does Features Like this Make Sense?",
  "url": "/competitions/amex-default-prediction/discussion/328508",
  "author_name": "PagalLadka",
  "post_date": "2022-06-01T15:20:46.604000",
  "votes": 2,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Inspired from Tabular Playground 2022 May Competition.</p>\n<p>train_test[0]['S_12_P_2'] = (train_test[0]['S_12'] + train_test[0]['P_2'] &gt; 1.4).astype(int) - (train_test[0]['S_12'] + train_test[0]['P_2'] &lt; 0.2).astype(int)\ntrain_test[1]['S_12_P_2'] = (train_test[1]['S_12'] + train_test[1]['P_2'] &gt; 1.4).astype(int) - (train_test[1]['S_12'] + train_test[1]['P_2'] &lt; 0.2).astype(int)</p>\n<p>Derived From :</p>\n<p>train['S_12'] = train.S_12.apply(lambda x : 1 if x &gt; 1 else x)<br>\ntrain['S_12_P_2'] = train['P_2'] + train['S_12'] <br>\nsns.pairplot(train[['S_12_P_2','target']][::10], hue =\"target\", palette ='coolwarm').fig.set_size_inches(15,15)</p>\n<p>Do this feature and this approach make sense?<br>\nIf not, Kindly Elaborate.!</p>\n<p>Sorry for the Inconvenience of me not posting a graph pic as Kaggle has disabled the feature for the Time being.</p>\n<p>Regards,<br>\nAnurag Mukherjee</p>",
  "messages": [
    {
      "id": 1808162,
      "postDate": "2022-06-01T15:20:46.603Z",
      "content": "<p>Inspired from Tabular Playground 2022 May Competition.</p>\n<p>train_test[0]['S_12_P_2'] = (train_test[0]['S_12'] + train_test[0]['P_2'] &gt; 1.4).astype(int) - (train_test[0]['S_12'] + train_test[0]['P_2'] &lt; 0.2).astype(int)\ntrain_test[1]['S_12_P_2'] = (train_test[1]['S_12'] + train_test[1]['P_2'] &gt; 1.4).astype(int) - (train_test[1]['S_12'] + train_test[1]['P_2'] &lt; 0.2).astype(int)</p>\n<p>Derived From :</p>\n<p>train['S_12'] = train.S_12.apply(lambda x : 1 if x &gt; 1 else x)<br>\ntrain['S_12_P_2'] = train['P_2'] + train['S_12'] <br>\nsns.pairplot(train[['S_12_P_2','target']][::10], hue =\"target\", palette ='coolwarm').fig.set_size_inches(15,15)</p>\n<p>Do this feature and this approach make sense?<br>\nIf not, Kindly Elaborate.!</p>\n<p>Sorry for the Inconvenience of me not posting a graph pic as Kaggle has disabled the feature for the Time being.</p>\n<p>Regards,<br>\nAnurag Mukherjee</p>",
      "rawMarkdown": "Inspired from Tabular Playground 2022 May Competition.\n\ntrain_test[0]['S_12_P_2'] = (train_test[0]['S_12'] + train_test[0]['P_2'] > 1.4).astype(int) - (train_test[0]['S_12'] + train_test[0]['P_2'] < 0.2).astype(int)\ntrain_test[1]['S_12_P_2'] = (train_test[1]['S_12'] + train_test[1]['P_2'] > 1.4).astype(int) - (train_test[1]['S_12'] + train_test[1]['P_2'] < 0.2).astype(int)\n\nDerived From :\n\ntrain['S_12'] = train.S_12.apply(lambda x : 1 if x > 1 else x)\ntrain['S_12_P_2'] = train['P_2'] + train['S_12'] \nsns.pairplot(train[['S_12_P_2','target']][::10], hue =\"target\", palette ='coolwarm').fig.set_size_inches(15,15)\n\nDo this feature and this approach make sense?\nIf not, Kindly Elaborate.!\n\nSorry for the Inconvenience of me not posting a graph pic as Kaggle has disabled the feature for the Time being.\n\nRegards,\nAnurag Mukherjee\n\n\n\n",
      "votes": 2
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1808162": "Inspired from Tabular Playground 2022 May Competition.\n\ntrain_test[0]['S_12_P_2'] = (train_test[0]['S_12'] + train_test[0]['P_2'] > 1.4).astype(int) - (train_test[0]['S_12'] + train_test[0]['P_2'] < 0.2).astype(int)\ntrain_test[1]['S_12_P_2'] = (train_test[1]['S_12'] + train_test[1]['P_2'] > 1.4).astype(int) - (train_test[1]['S_12'] + train_test[1]['P_2'] < 0.2).astype(int)\n\nDerived From :\n\ntrain['S_12'] = train.S_12.apply(lambda x : 1 if x > 1 else x)\ntrain['S_12_P_2'] = train['P_2'] + train['S_12'] \nsns.pairplot(train[['S_12_P_2','target']][::10], hue =\"target\", palette ='coolwarm').fig.set_size_inches(15,15)\n\nDo this feature and this approach make sense?\nIf not, Kindly Elaborate.!\n\nSorry for the Inconvenience of me not posting a graph pic as Kaggle has disabled the feature for the Time being.\n\nRegards,\nAnurag Mukherjee\n\n\n\n"
  }
}