{
  "id": 130241,
  "title": "Indeed,smoothing  helps target encoding",
  "url": "/competitions/cat-in-the-dat-ii/discussion/130241",
  "author_name": "",
  "post_date": "2020-02-12T23:45:23.369299700Z",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>xgboost, lossguide:</p>\n\n<p>smooth12 \nLB:  0.78607, \nCV:   (0.789341+0.788633+0.787345+0.787166+0.788660)/5=0.788229\nst12 <br>\nLB:  0.78599, \nCV: (0.788576+0.787062+0.787425+0.788529+0.789215)/5=0.7881614</p>\n\n<p>Formula:\n<a href=\"https://maxhalford.github.io/blog/target-encoding-done-the-right-way/\">https://maxhalford.github.io/blog/target-encoding-done-the-right-way/</a></p>",
  "messages": [
    {
      "id": "744560",
      "postDate": "02/12/2020 23:45:23",
      "content": "<p>xgboost, lossguide:</p>\n\n<p>smooth12 \nLB:  0.78607, \nCV:   (0.789341+0.788633+0.787345+0.787166+0.788660)/5=0.788229\nst12 <br>\nLB:  0.78599, \nCV: (0.788576+0.787062+0.787425+0.788529+0.789215)/5=0.7881614</p>\n\n<p>Formula:\n<a href=\"https://maxhalford.github.io/blog/target-encoding-done-the-right-way/\">https://maxhalford.github.io/blog/target-encoding-done-the-right-way/</a></p>",
      "rawMarkdown": "xgboost, lossguide:\n\n\nsmooth12 \nLB:  0.78607, \nCV:   (0.789341+0.788633+0.787345+0.787166+0.788660)/5=0.788229\nst12            \nLB:  0.78599, \nCV: (0.788576+0.787062+0.787425+0.788529+0.789215)/5=0.7881614\n\n\n\nFormula:\nhttps://maxhalford.github.io/blog/target-encoding-done-the-right-way/",
      "votes": null
    },
    {
      "id": "744611",
      "postDate": "02/13/2020 01:18:28",
      "content": "<p>hmm, so you one hot encoded the targeting encoding for XGBoost?</p>",
      "rawMarkdown": "hmm, so you one hot encoded the targeting encoding for XGBoost?",
      "votes": null
    },
    {
      "id": "744694",
      "postDate": "02/13/2020 03:27:12",
      "content": "<p>No. target encoding,plus smoothing</p>",
      "rawMarkdown": "No. target encoding,plus smoothing",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 744611,
      "author_name": "rkotwani",
      "author_url": "",
      "post_date": "02/13/2020 01:18:28",
      "content": "<p>hmm, so you one hot encoded the targeting encoding for XGBoost?</p>",
      "votes": null,
      "replies": [
        {
          "id": 744694,
          "author_name": "kailai",
          "author_url": "",
          "post_date": "02/13/2020 03:27:12",
          "content": "<p>No. target encoding,plus smoothing</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "744560": "xgboost, lossguide:\n\n\nsmooth12 \nLB:  0.78607, \nCV:   (0.789341+0.788633+0.787345+0.787166+0.788660)/5=0.788229\nst12            \nLB:  0.78599, \nCV: (0.788576+0.787062+0.787425+0.788529+0.789215)/5=0.7881614\n\n\n\nFormula:\nhttps://maxhalford.github.io/blog/target-encoding-done-the-right-way/",
    "744611": "hmm, so you one hot encoded the targeting encoding for XGBoost?",
    "744694": "No. target encoding,plus smoothing"
  },
  "source": "meta"
}