{
  "id": 241155,
  "title": "Has ensembling helped any one ?",
  "url": "/competitions/bms-molecular-translation/discussion/241155",
  "author_name": "",
  "post_date": "2021-05-23T09:57:15.484274100Z",
  "votes": 5,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Has anyone tried ensembling. And if yes, can anyone share some insights on how it worked and the strategies used for the same ?</p>",
  "messages": [
    {
      "id": "1319544",
      "postDate": "05/23/2021 09:57:15",
      "content": "<p>Has anyone tried ensembling. And if yes, can anyone share some insights on how it worked and the strategies used for the same ?</p>",
      "rawMarkdown": "Has anyone tried ensembling. And if yes, can anyone share some insights on how it worked and the strategies used for the same ?",
      "votes": null
    },
    {
      "id": "1323413",
      "postDate": "05/26/2021 08:04:54",
      "content": "<p>Yes, it has helped. 😃</p>",
      "rawMarkdown": "Yes, it has helped. 😃",
      "votes": null
    },
    {
      "id": "1324861",
      "postDate": "05/27/2021 09:36:41",
      "content": "<p>Specifically, I just used some models I had previously discarded as \"poor\" and got a 0.30 improvement in LB score over my best single model. </p>",
      "rawMarkdown": "Specifically, I just used some models I had previously discarded as \"poor\" and got a 0.30 improvement in LB score over my best single model.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1323413,
      "author_name": "talktocharles",
      "author_url": "",
      "post_date": "05/26/2021 08:04:54",
      "content": "<p>Yes, it has helped. 😃</p>",
      "votes": null,
      "replies": [
        {
          "id": 1324861,
          "author_name": "talktocharles",
          "author_url": "",
          "post_date": "05/27/2021 09:36:41",
          "content": "<p>Specifically, I just used some models I had previously discarded as \"poor\" and got a 0.30 improvement in LB score over my best single model. </p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1319544": "Has anyone tried ensembling. And if yes, can anyone share some insights on how it worked and the strategies used for the same ?",
    "1323413": "Yes, it has helped. 😃",
    "1324861": "Specifically, I just used some models I had previously discarded as \"poor\" and got a 0.30 improvement in LB score over my best single model."
  },
  "source": "meta"
}