{
  "id": 74152,
  "title": "Meta Features And Correlation Heatmap",
  "url": "/competitions/quora-insincere-questions-classification/discussion/74152",
  "author_name": "Shahebaz Mohammad",
  "post_date": "2018-12-09T09:21:36.478000",
  "votes": 21,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Apart from embeddings and updating them. Here are few ways you can use meta-features. </p>\n\n<ol>\n<li>Logistic/XGB Classification Model on Meta-features</li>\n<li>Concatenate them as input in Neural Networks</li>\n<li>Use them in your level 2/3 stacking of models. </li>\n</ol>\n\n<p>Few handy features I am experimenting with are;\n1. Length\n2. Num words/chars/\n3. Emoji/Smileys\n4. POS tagged features - Noun/Adverb/Adjective and their ratios\n5. No. of toxic words </p>\n\n<p>Here's an implemented <a href=\"https://www.kaggle.com/shaz13/feature-engineering-for-nlp-classification/notebook\">kernel on Feature Engineering For NLP Classification\n</a> </p>\n\n<h2>Correlation heat map for meta features</h2>\n\n<p><img src=\"https://i.ibb.co/QHqM3wG/results-28-0.png\" alt=\"Coor Heatmap\"></p>\n\n<p><a href=\"https://ibb.co/z2LZTvw\">Link to HD Image</a></p>",
  "messages": [
    {
      "id": 436001,
      "postDate": "2018-12-09T09:21:36.477Z",
      "content": "<p>Apart from embeddings and updating them. Here are few ways you can use meta-features. </p>\n\n<ol>\n<li>Logistic/XGB Classification Model on Meta-features</li>\n<li>Concatenate them as input in Neural Networks</li>\n<li>Use them in your level 2/3 stacking of models. </li>\n</ol>\n\n<p>Few handy features I am experimenting with are;\n1. Length\n2. Num words/chars/\n3. Emoji/Smileys\n4. POS tagged features - Noun/Adverb/Adjective and their ratios\n5. No. of toxic words </p>\n\n<p>Here's an implemented <a href=\"https://www.kaggle.com/shaz13/feature-engineering-for-nlp-classification/notebook\">kernel on Feature Engineering For NLP Classification\n</a> </p>\n\n<h2>Correlation heat map for meta features</h2>\n\n<p><img src=\"https://i.ibb.co/QHqM3wG/results-28-0.png\" alt=\"Coor Heatmap\"></p>\n\n<p><a href=\"https://ibb.co/z2LZTvw\">Link to HD Image</a></p>",
      "rawMarkdown": "Apart from embeddings and updating them. Here are few ways you can use meta-features. \n\n1. Logistic/XGB Classification Model on Meta-features\n2. Concatenate them as input in Neural Networks\n3. Use them in your level 2/3 stacking of models. \n\nFew handy features I am experimenting with are;\n1. Length\n2. Num words/chars/\n3. Emoji/Smileys\n4. POS tagged features - Noun/Adverb/Adjective and their ratios\n5. No. of toxic words \n\nHere's an implemented [kernel on Feature Engineering For NLP Classification\n][1] \n\n## Correlation heat map for meta features\n\n![Coor Heatmap][3]\n\n [Link to HD Image][2]\n\n  [1]: https://www.kaggle.com/shaz13/feature-engineering-for-nlp-classification/notebook\n  [2]: https://ibb.co/z2LZTvw\n  [3]: https://i.ibb.co/QHqM3wG/results-28-0.png",
      "votes": 20
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "436001": "Apart from embeddings and updating them. Here are few ways you can use meta-features. \n\n1. Logistic/XGB Classification Model on Meta-features\n2. Concatenate them as input in Neural Networks\n3. Use them in your level 2/3 stacking of models. \n\nFew handy features I am experimenting with are;\n1. Length\n2. Num words/chars/\n3. Emoji/Smileys\n4. POS tagged features - Noun/Adverb/Adjective and their ratios\n5. No. of toxic words \n\nHere's an implemented [kernel on Feature Engineering For NLP Classification\n][1] \n\n## Correlation heat map for meta features\n\n![Coor Heatmap][3]\n\n [Link to HD Image][2]\n\n  [1]: https://www.kaggle.com/shaz13/feature-engineering-for-nlp-classification/notebook\n  [2]: https://ibb.co/z2LZTvw\n  [3]: https://i.ibb.co/QHqM3wG/results-28-0.png"
  }
}