{
  "id": 437510,
  "title": "Silver Medal  with a Simple Constant : 0.99",
  "url": "/competitions/rsna-2023-abdominal-trauma-detection/discussion/437510",
  "author_name": "justin1357",
  "post_date": "2023-09-07T04:06:16.770000",
  "votes": 6,
  "comment_count": 5,
  "views": 0,
  "content": "<p>Surprising as it may seem, setting the prediction value for a specific label 'extravasation_injury' to a constant 0.99 yielded outstanding results in my prediction. Could high scores on the LB simply be masking an overfit model? </p>",
  "messages": [
    {
      "id": 2427033,
      "postDate": "2023-09-07T04:06:16.770Z",
      "content": "<p>Surprising as it may seem, setting the prediction value for a specific label 'extravasation_injury' to a constant 0.99 yielded outstanding results in my prediction. Could high scores on the LB simply be masking an overfit model? </p>",
      "rawMarkdown": "Surprising as it may seem, setting the prediction value for a specific label 'extravasation_injury' to a constant 0.99 yielded outstanding results in my prediction. Could high scores on the LB simply be masking an overfit model? ",
      "votes": 6
    },
    {
      "id": 2429633,
      "postDate": "2023-09-08T17:32:36.657Z",
      "content": "<p>Given the number of issues with the notebook and dataset that have been discovered, this doesn't really surprise me too much. Hopefully the hidden dataset crushes the constant so it doesn't actually affect the competition, but good discovery nonetheless!</p>",
      "rawMarkdown": "Given the number of issues with the notebook and dataset that have been discovered, this doesn't really surprise me too much. Hopefully the hidden dataset crushes the constant so it doesn't actually affect the competition, but good discovery nonetheless!",
      "votes": 1,
      "replies": [
        {
          "id": 2429637,
          "postDate": "2023-09-08T17:34:56.750Z",
          "content": "<p>Thanks a lot!</p>",
          "rawMarkdown": "Thanks a lot!",
          "votes": -1
        }
      ]
    },
    {
      "id": 2427146,
      "postDate": "2023-09-07T05:29:28.850Z",
      "content": "<p>Expecting a serious shakeup…</p>",
      "rawMarkdown": "Expecting a serious shakeup..."
    },
    {
      "id": 2427345,
      "postDate": "2023-09-07T07:08:40.037Z",
      "rawMarkdown": "",
      "votes": -2,
      "isDeleted": true,
      "replies": [
        {
          "id": 2427445,
          "postDate": "2023-09-07T08:18:46.580Z",
          "content": "<p>Thank you for your detailed reply ! I'll try to find solutions.(though polished by chatGPT haha)</p>",
          "rawMarkdown": "Thank you for your detailed reply ! I'll try to find solutions.(though polished by chatGPT haha)"
        }
      ]
    }
  ],
  "comments": [
    {
      "id": 2429633,
      "author_name": "Andrew Scott",
      "author_url": "",
      "post_date": "2023-09-08T17:32:36.657000",
      "content": "<p>Given the number of issues with the notebook and dataset that have been discovered, this doesn't really surprise me too much. Hopefully the hidden dataset crushes the constant so it doesn't actually affect the competition, but good discovery nonetheless!</p>",
      "votes": 1,
      "replies": [
        {
          "id": 2429637,
          "author_name": "Justin6499",
          "author_url": "",
          "post_date": "2023-09-08T17:34:56.750000",
          "content": "<p>Thanks a lot!</p>",
          "votes": -1,
          "replies": []
        }
      ]
    },
    {
      "id": 2427146,
      "author_name": "Justin6499",
      "author_url": "",
      "post_date": "2023-09-07T05:29:28.850000",
      "content": "<p>Expecting a serious shakeup…</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 2427345,
      "author_name": "",
      "author_url": "",
      "post_date": "2023-09-07T07:08:40.037000",
      "content": "",
      "votes": -2,
      "replies": [
        {
          "id": 2427445,
          "author_name": "Justin6499",
          "author_url": "",
          "post_date": "2023-09-07T08:18:46.580000",
          "content": "<p>Thank you for your detailed reply ! I'll try to find solutions.(though polished by chatGPT haha)</p>",
          "votes": 0,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2427033": "Surprising as it may seem, setting the prediction value for a specific label 'extravasation_injury' to a constant 0.99 yielded outstanding results in my prediction. Could high scores on the LB simply be masking an overfit model? ",
    "2429633": "Given the number of issues with the notebook and dataset that have been discovered, this doesn't really surprise me too much. Hopefully the hidden dataset crushes the constant so it doesn't actually affect the competition, but good discovery nonetheless!",
    "2427146": "Expecting a serious shakeup...",
    "2427345": ""
  }
}