{
  "id": 344619,
  "title": "how does rank ensemble work?",
  "url": "/competitions/amex-default-prediction/discussion/344619",
  "author_name": "",
  "post_date": "2022-08-16T01:09:46.819046900Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>i saw several public notebooks used rank ensemble. i am not familiar with this ensemble method. who can give a detailed explanation? thanks a lot. </p>\n<p>for df in dfs:<br>\n    submit['prediction'] += rankdata(df['prediction'])/df.shape[0]</p>",
  "messages": [
    {
      "id": "1900371",
      "postDate": "08/16/2022 01:09:46",
      "content": "<p>i saw several public notebooks used rank ensemble. i am not familiar with this ensemble method. who can give a detailed explanation? thanks a lot. </p>\n<p>for df in dfs:<br>\n    submit['prediction'] += rankdata(df['prediction'])/df.shape[0]</p>",
      "rawMarkdown": "i saw several public notebooks used rank ensemble. i am not familiar with this ensemble method. who can give a detailed explanation? thanks a lot. \n\nfor df in dfs:\n    submit['prediction'] += rankdata(df['prediction'])/df.shape[0]",
      "votes": null
    },
    {
      "id": "1900401",
      "postDate": "08/16/2022 02:01:45",
      "content": "<p>this competition's metric is about auc, and auc is a metric to evalute the rank of every sample, so this ensemble method is efficient.</p>",
      "rawMarkdown": "this competition's metric is about auc, and auc is a metric to evalute the rank of every sample, so this ensemble method is efficient.",
      "votes": null
    },
    {
      "id": "1900670",
      "postDate": "08/16/2022 07:25:26",
      "content": "<p>thanks for nice explanation!</p>",
      "rawMarkdown": "thanks for nice explanation!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1900401,
      "author_name": "brilliw",
      "author_url": "",
      "post_date": "08/16/2022 02:01:45",
      "content": "<p>this competition's metric is about auc, and auc is a metric to evalute the rank of every sample, so this ensemble method is efficient.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1900670,
          "author_name": "shuozhang",
          "author_url": "",
          "post_date": "08/16/2022 07:25:26",
          "content": "<p>thanks for nice explanation!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1900371": "i saw several public notebooks used rank ensemble. i am not familiar with this ensemble method. who can give a detailed explanation? thanks a lot. \n\nfor df in dfs:\n    submit['prediction'] += rankdata(df['prediction'])/df.shape[0]",
    "1900401": "this competition's metric is about auc, and auc is a metric to evalute the rank of every sample, so this ensemble method is efficient.",
    "1900670": "thanks for nice explanation!"
  },
  "source": "meta"
}