{
  "id": 342024,
  "title": "Look for team member and share my model",
  "url": "/competitions/amex-default-prediction/discussion/342024",
  "author_name": "Alger, Yang",
  "post_date": "2022-08-05T09:04:02.243000",
  "votes": 5,
  "comment_count": 5,
  "views": 0,
  "content": "<p>My best model is catboost on GPU, and only used 38x features to get CV: 0.79601 vs LB:0.797…</p>\n<p>I try RFE and Null importance to do feature selection from 1365 --&gt; 38x </p>\n<p>After low features , i try to tuning parameters by optuna and get 0.797 on LB</p>\n<p>If someone feel my model can help you, let me know to build a team ~~ or let me join </p>\n<p>THX❤️</p>",
  "messages": [
    {
      "id": 1885573,
      "postDate": "2022-08-05T09:04:02.243Z",
      "content": "<p>My best model is catboost on GPU, and only used 38x features to get CV: 0.79601 vs LB:0.797…</p>\n<p>I try RFE and Null importance to do feature selection from 1365 --&gt; 38x </p>\n<p>After low features , i try to tuning parameters by optuna and get 0.797 on LB</p>\n<p>If someone feel my model can help you, let me know to build a team ~~ or let me join </p>\n<p>THX❤️</p>",
      "rawMarkdown": "My best model is catboost on GPU, and only used 38x features to get CV: 0.79601 vs LB:0.797...\n\nI try RFE and Null importance to do feature selection from 1365 --> 38x \n\nAfter low features , i try to tuning parameters by optuna and get 0.797 on LB\n\nIf someone feel my model can help you, let me know to build a team ~~ or let me join \n\nTHX❤️",
      "votes": 5
    },
    {
      "id": 1886251,
      "postDate": "2022-08-05T16:48:14.320Z",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/algeryang\" target=\"_blank\">@algeryang</a> great score with catboost. I have a doubt, what is RCF?</p>",
      "rawMarkdown": "Hi @algeryang great score with catboost. I have a doubt, what is RCF?",
      "votes": 1,
      "replies": [
        {
          "id": 1886582,
          "postDate": "2022-08-06T02:13:33.977Z",
          "content": "<p>sorry, it's Recursive feature elimination ~~<br>\nRFE</p>",
          "rawMarkdown": "sorry, it's Recursive feature elimination ~~\nRFE",
          "votes": 1
        }
      ]
    },
    {
      "id": 1886411,
      "postDate": "2022-08-05T19:36:02.293Z",
      "content": "<p>Sorry if this is a dumb question but are you saying you use 38 features or 380-390 features (x is a mystery)?</p>\n<p>Either way nice work man.</p>",
      "rawMarkdown": "Sorry if this is a dumb question but are you saying you use 38 features or 380-390 features (x is a mystery)?\n\nEither way nice work man.",
      "replies": [
        {
          "id": 1886584,
          "postDate": "2022-08-06T02:14:15.727Z",
          "content": "<p>380 ~ 390 </p>",
          "rawMarkdown": "380 ~ 390 "
        }
      ]
    },
    {
      "id": 1885840,
      "postDate": "2022-08-05T12:02:21.937Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 1886251,
      "author_name": "AKR",
      "author_url": "",
      "post_date": "2022-08-05T16:48:14.320000",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/algeryang\" target=\"_blank\">@algeryang</a> great score with catboost. I have a doubt, what is RCF?</p>",
      "votes": 1,
      "replies": [
        {
          "id": 1886582,
          "author_name": "Alger, Yang",
          "author_url": "",
          "post_date": "2022-08-06T02:13:33.977000",
          "content": "<p>sorry, it's Recursive feature elimination ~~<br>\nRFE</p>",
          "votes": 1,
          "replies": []
        }
      ]
    },
    {
      "id": 1886411,
      "author_name": "S Charlesworth",
      "author_url": "",
      "post_date": "2022-08-05T19:36:02.293000",
      "content": "<p>Sorry if this is a dumb question but are you saying you use 38 features or 380-390 features (x is a mystery)?</p>\n<p>Either way nice work man.</p>",
      "votes": 0,
      "replies": [
        {
          "id": 1886584,
          "author_name": "Alger, Yang",
          "author_url": "",
          "post_date": "2022-08-06T02:14:15.727000",
          "content": "<p>380 ~ 390 </p>",
          "votes": 0,
          "replies": []
        }
      ]
    },
    {
      "id": 1885840,
      "author_name": "",
      "author_url": "",
      "post_date": "2022-08-05T12:02:21.937000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1885573": "My best model is catboost on GPU, and only used 38x features to get CV: 0.79601 vs LB:0.797...\n\nI try RFE and Null importance to do feature selection from 1365 --> 38x \n\nAfter low features , i try to tuning parameters by optuna and get 0.797 on LB\n\nIf someone feel my model can help you, let me know to build a team ~~ or let me join \n\nTHX❤️",
    "1886251": "Hi @algeryang great score with catboost. I have a doubt, what is RCF?",
    "1886411": "Sorry if this is a dumb question but are you saying you use 38 features or 380-390 features (x is a mystery)?\n\nEither way nice work man.",
    "1885840": ""
  }
}