{
  "id": 128403,
  "title": "lb goes with cv",
  "url": "/competitions/cat-in-the-dat-ii/discussion/128403",
  "author_name": "",
  "post_date": "2020-01-31T03:40:32.380226800Z",
  "votes": 3,
  "comment_count": 6,
  "views": 0,
  "content": "<p>xgb,lossguide,5 folds,oof : </p>\n\n<p>(0.786930+0.789511+0.787825+0.789248+0.787455)/5=0.7881938,  lb 0.78600\n(0.789183+0.788403+0.787401+0.787077+0.788675)/5=0.7881478,  lb 0.78599\n(0.787018+0.789554+0.787731+0.789147+0.787276)/5=0.7881452,   lb 0.78599\n(0.787370+0.789144+0.787799+0.789487+0.786798)/5=0.7881196,  lb 0.78596</p>",
  "messages": [
    {
      "id": "733377",
      "postDate": "01/31/2020 03:40:32",
      "content": "<p>xgb,lossguide,5 folds,oof : </p>\n\n<p>(0.786930+0.789511+0.787825+0.789248+0.787455)/5=0.7881938,  lb 0.78600\n(0.789183+0.788403+0.787401+0.787077+0.788675)/5=0.7881478,  lb 0.78599\n(0.787018+0.789554+0.787731+0.789147+0.787276)/5=0.7881452,   lb 0.78599\n(0.787370+0.789144+0.787799+0.789487+0.786798)/5=0.7881196,  lb 0.78596</p>",
      "rawMarkdown": "xgb,lossguide,5 folds,oof : \n\n(0.786930+0.789511+0.787825+0.789248+0.787455)/5=0.7881938,  lb 0.78600\n(0.789183+0.788403+0.787401+0.787077+0.788675)/5=0.7881478,  lb 0.78599\n(0.787018+0.789554+0.787731+0.789147+0.787276)/5=0.7881452,   lb 0.78599\n(0.787370+0.789144+0.787799+0.789487+0.786798)/5=0.7881196,  lb 0.78596",
      "votes": null
    },
    {
      "id": "734662",
      "postDate": "02/01/2020 19:55:55",
      "content": "<p>It seems that Xgb might not be the right way...</p>\n\n<p>And I don’t know the right way neither.</p>",
      "rawMarkdown": "It seems that Xgb might not be the right way...\n\nAnd I don’t know the right way neither.",
      "votes": null
    },
    {
      "id": "734830",
      "postDate": "02/02/2020 04:06:51",
      "content": "<p>Thank you for comment.You are right about model selection!</p>",
      "rawMarkdown": "Thank you for comment.You are right about model selection!",
      "votes": null
    },
    {
      "id": "735308",
      "postDate": "02/02/2020 21:21:44",
      "content": "<p>Didn't have much luck with xgb either. Basic 2 - 3 layer NN works okay. Not sure what models the top solutions are using.</p>",
      "rawMarkdown": "Didn't have much luck with xgb either. Basic 2 - 3 layer NN works okay. Not sure what models the top solutions are using.",
      "votes": null
    },
    {
      "id": "735427",
      "postDate": "02/03/2020 02:50:00",
      "content": "<p>Maybe,plus swap noises?</p>",
      "rawMarkdown": "Maybe,plus swap noises?",
      "votes": null
    },
    {
      "id": "736154",
      "postDate": "02/03/2020 22:25:44",
      "content": "<p>You could check out the kernels to get an idea. I have published one which might help you.</p>",
      "rawMarkdown": "You could check out the kernels to get an idea. I have published one which might help you.",
      "votes": null
    },
    {
      "id": "737483",
      "postDate": "02/05/2020 11:57:23",
      "content": "<p>CatBoost worked the best for me, until now. </p>",
      "rawMarkdown": "CatBoost worked the best for me, until now.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 734662,
      "author_name": "pfichou",
      "author_url": "",
      "post_date": "02/01/2020 19:55:55",
      "content": "<p>It seems that Xgb might not be the right way...</p>\n\n<p>And I don’t know the right way neither.</p>",
      "votes": null,
      "replies": [
        {
          "id": 734830,
          "author_name": "kailai",
          "author_url": "",
          "post_date": "02/02/2020 04:06:51",
          "content": "<p>Thank you for comment.You are right about model selection!</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 735308,
          "author_name": "advaithasabnis",
          "author_url": "",
          "post_date": "02/02/2020 21:21:44",
          "content": "<p>Didn't have much luck with xgb either. Basic 2 - 3 layer NN works okay. Not sure what models the top solutions are using.</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 735427,
          "author_name": "kailai",
          "author_url": "",
          "post_date": "02/03/2020 02:50:00",
          "content": "<p>Maybe,plus swap noises?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 737483,
          "author_name": "catadanna",
          "author_url": "",
          "post_date": "02/05/2020 11:57:23",
          "content": "<p>CatBoost worked the best for me, until now. </p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 736154,
      "author_name": "frankmollard",
      "author_url": "",
      "post_date": "02/03/2020 22:25:44",
      "content": "<p>You could check out the kernels to get an idea. I have published one which might help you.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "733377": "xgb,lossguide,5 folds,oof : \n\n(0.786930+0.789511+0.787825+0.789248+0.787455)/5=0.7881938,  lb 0.78600\n(0.789183+0.788403+0.787401+0.787077+0.788675)/5=0.7881478,  lb 0.78599\n(0.787018+0.789554+0.787731+0.789147+0.787276)/5=0.7881452,   lb 0.78599\n(0.787370+0.789144+0.787799+0.789487+0.786798)/5=0.7881196,  lb 0.78596",
    "734662": "It seems that Xgb might not be the right way...\n\nAnd I don’t know the right way neither.",
    "734830": "Thank you for comment.You are right about model selection!",
    "735308": "Didn't have much luck with xgb either. Basic 2 - 3 layer NN works okay. Not sure what models the top solutions are using.",
    "735427": "Maybe,plus swap noises?",
    "736154": "You could check out the kernels to get an idea. I have published one which might help you.",
    "737483": "CatBoost worked the best for me, until now."
  },
  "source": "meta"
}