{
  "id": 335705,
  "title": "Does the target encoding work?",
  "url": "/competitions/amex-default-prediction/discussion/335705",
  "author_name": "",
  "post_date": "2022-07-07T13:08:36.716546600Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Our team tried CatboostEncoder to make new features, but It doesn't work well ( If I had to say, minor CV score improvement)<br>\nDid you try target encoding which improves CV or public score?</p>",
  "messages": [
    {
      "id": "1846915",
      "postDate": "07/07/2022 13:08:36",
      "content": "<p>Our team tried CatboostEncoder to make new features, but It doesn't work well ( If I had to say, minor CV score improvement)<br>\nDid you try target encoding which improves CV or public score?</p>",
      "rawMarkdown": "Our team tried CatboostEncoder to make new features, but It doesn't work well ( If I had to say, minor CV score improvement)\nDid you try target encoding which improves CV or public score?",
      "votes": null
    },
    {
      "id": "1847484",
      "postDate": "07/08/2022 00:49:07",
      "content": "<p>Didn't try it yet.<br>\nIn my personal experience I have found that target encoding generally works better than one-hot encoding, especially if you have a lot of categorical variables. I would recommend trying it out on your data to see if you can improve your results.</p>",
      "rawMarkdown": "Didn't try it yet.\nIn my personal experience I have found that target encoding generally works better than one-hot encoding, especially if you have a lot of categorical variables. I would recommend trying it out on your data to see if you can improve your results.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1847484,
      "author_name": "thedevastator",
      "author_url": "",
      "post_date": "07/08/2022 00:49:07",
      "content": "<p>Didn't try it yet.<br>\nIn my personal experience I have found that target encoding generally works better than one-hot encoding, especially if you have a lot of categorical variables. I would recommend trying it out on your data to see if you can improve your results.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1846915": "Our team tried CatboostEncoder to make new features, but It doesn't work well ( If I had to say, minor CV score improvement)\nDid you try target encoding which improves CV or public score?",
    "1847484": "Didn't try it yet.\nIn my personal experience I have found that target encoding generally works better than one-hot encoding, especially if you have a lot of categorical variables. I would recommend trying it out on your data to see if you can improve your results."
  },
  "source": "meta"
}