{
  "id": 375404,
  "title": "Basic feature engineering,hope it helps.",
  "url": "/competitions/nfl-player-contact-detection/discussion/375404",
  "author_name": "",
  "post_date": "2023-01-01T11:53:52.151167200Z",
  "votes": 2,
  "comment_count": 5,
  "views": 0,
  "content": "<p><a href=\"https://www.kaggle.com/competitions/ieee-fraud-detection/discussion/108575\" target=\"_blank\">discussion</a></p>",
  "messages": [
    {
      "id": "2082283",
      "postDate": "01/01/2023 11:53:52",
      "content": "<p><a href=\"https://www.kaggle.com/competitions/ieee-fraud-detection/discussion/108575\" target=\"_blank\">discussion</a></p>",
      "rawMarkdown": "[discussion](https://www.kaggle.com/competitions/ieee-fraud-detection/discussion/108575)",
      "votes": null
    },
    {
      "id": "2096625",
      "postDate": "01/12/2023 07:10:13",
      "content": "<p>Can you please share some more information about your work? </p>\n<p>The Devastator.</p>",
      "rawMarkdown": "Can you please share some more information about your work? \n\nThe Devastator.",
      "votes": null
    },
    {
      "id": "2099664",
      "postDate": "01/14/2023 16:10:45",
      "content": "<p>If my lb score reaches 0.7, I will share more work.</p>",
      "rawMarkdown": "If my lb score reaches 0.7, I will share more work.",
      "votes": null
    },
    {
      "id": "2121068",
      "postDate": "01/30/2023 03:28:47",
      "content": "<p>Features(div and sub) are very useful for tree model in this competetion.This is an example:train['x_div'] = train['x_1']/train['x_2'],and this:['x_sub'] = train['x_1']-train['x_2'].</p>",
      "rawMarkdown": "Features(div and sub) are very useful for tree model in this competetion.This is an example:train['x_div'] = train['x_1']/train['x_2'],and this:['x_sub'] = train['x_1']-train['x_2'].",
      "votes": null
    },
    {
      "id": "2121579",
      "postDate": "01/30/2023 12:13:25",
      "content": "<p>I'd recommend being careful with features like train['x_1']/train['x_2'], sub may be useful, but the importance of features like div is most likely accidental and may not generalize well. If I think a feature should not have an impact or does not make sense, I usually prefer to exclude it.</p>",
      "rawMarkdown": "I'd recommend being careful with features like train['x_1']/train['x_2'], sub may be useful, but the importance of features like div is most likely accidental and may not generalize well. If I think a feature should not have an impact or does not make sense, I usually prefer to exclude it.",
      "votes": null
    },
    {
      "id": "2156713",
      "postDate": "02/23/2023 13:58:20",
      "content": "<p>appreciate</p>",
      "rawMarkdown": "appreciate",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2096625,
      "author_name": "thedevastator",
      "author_url": "",
      "post_date": "01/12/2023 07:10:13",
      "content": "<p>Can you please share some more information about your work? </p>\n<p>The Devastator.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2099664,
          "author_name": "cjzccc",
          "author_url": "",
          "post_date": "01/14/2023 16:10:45",
          "content": "<p>If my lb score reaches 0.7, I will share more work.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2121068,
      "author_name": "cjzccc",
      "author_url": "",
      "post_date": "01/30/2023 03:28:47",
      "content": "<p>Features(div and sub) are very useful for tree model in this competetion.This is an example:train['x_div'] = train['x_1']/train['x_2'],and this:['x_sub'] = train['x_1']-train['x_2'].</p>",
      "votes": null,
      "replies": [
        {
          "id": 2121579,
          "author_name": "dmytropoplavskiy",
          "author_url": "",
          "post_date": "01/30/2023 12:13:25",
          "content": "<p>I'd recommend being careful with features like train['x_1']/train['x_2'], sub may be useful, but the importance of features like div is most likely accidental and may not generalize well. If I think a feature should not have an impact or does not make sense, I usually prefer to exclude it.</p>",
          "votes": null,
          "replies": [
            {
              "id": 2156713,
              "author_name": "cjzccc",
              "author_url": "",
              "post_date": "02/23/2023 13:58:20",
              "content": "<p>appreciate</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2082283": "[discussion](https://www.kaggle.com/competitions/ieee-fraud-detection/discussion/108575)",
    "2096625": "Can you please share some more information about your work? \n\nThe Devastator.",
    "2099664": "If my lb score reaches 0.7, I will share more work.",
    "2121068": "Features(div and sub) are very useful for tree model in this competetion.This is an example:train['x_div'] = train['x_1']/train['x_2'],and this:['x_sub'] = train['x_1']-train['x_2'].",
    "2121579": "I'd recommend being careful with features like train['x_1']/train['x_2'], sub may be useful, but the importance of features like div is most likely accidental and may not generalize well. If I think a feature should not have an impact or does not make sense, I usually prefer to exclude it.",
    "2156713": "appreciate"
  },
  "source": "meta"
}