{
  "id": 508764,
  "title": "Tiny scoring dataset?",
  "url": "/competitions/rsna-2024-lumbar-spine-degenerative-classification/discussion/508764",
  "author_name": "",
  "post_date": "2024-05-31T01:50:38.080694100Z",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I was just wondering why the scoring dataset is so small, 25 rows of three predictions, which is tiny by Kaggle standards (and very likely to result in easy overfitting) is their a reason behind only using such a small fraction of the data for scoring?</p>",
  "messages": [
    {
      "id": "2846173",
      "postDate": "05/31/2024 01:50:38",
      "content": "<p>I was just wondering why the scoring dataset is so small, 25 rows of three predictions, which is tiny by Kaggle standards (and very likely to result in easy overfitting) is their a reason behind only using such a small fraction of the data for scoring?</p>",
      "rawMarkdown": "I was just wondering why the scoring dataset is so small, 25 rows of three predictions, which is tiny by Kaggle standards (and very likely to result in easy overfitting) is their a reason behind only using such a small fraction of the data for scoring?",
      "votes": null
    },
    {
      "id": "2846637",
      "postDate": "05/31/2024 07:21:56",
      "content": "<p>25 rows is not the full test set most of them is hidden, 25 rows just a sample, it means when you submit a notebook the hidden set will replace it.So you should take care about the full set. Here is the description in the data page:</p>\n<blockquote>\n  <p>This competition uses a hidden test. When your submitted notebook is scored, the actual test data (including a full length sample submission) will be made available to your notebook.</p>\n</blockquote>",
      "rawMarkdown": "25 rows is not the full test set most of them is hidden, 25 rows just a sample, it means when you submit a notebook the hidden set will replace it.So you should take care about the full set. Here is the description in the data page:\n>This competition uses a hidden test. When your submitted notebook is scored, the actual test data (including a full length sample submission) will be made available to your notebook.",
      "votes": null
    },
    {
      "id": "2846653",
      "postDate": "05/31/2024 07:31:35",
      "content": "<p>Thanks, useful comment and source of my problems!</p>",
      "rawMarkdown": "Thanks, useful comment and source of my problems!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2846637,
      "author_name": "athrunzala",
      "author_url": "",
      "post_date": "05/31/2024 07:21:56",
      "content": "<p>25 rows is not the full test set most of them is hidden, 25 rows just a sample, it means when you submit a notebook the hidden set will replace it.So you should take care about the full set. Here is the description in the data page:</p>\n<blockquote>\n  <p>This competition uses a hidden test. When your submitted notebook is scored, the actual test data (including a full length sample submission) will be made available to your notebook.</p>\n</blockquote>",
      "votes": null,
      "replies": [
        {
          "id": 2846653,
          "author_name": "qmot66",
          "author_url": "",
          "post_date": "05/31/2024 07:31:35",
          "content": "<p>Thanks, useful comment and source of my problems!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2846173": "I was just wondering why the scoring dataset is so small, 25 rows of three predictions, which is tiny by Kaggle standards (and very likely to result in easy overfitting) is their a reason behind only using such a small fraction of the data for scoring?",
    "2846637": "25 rows is not the full test set most of them is hidden, 25 rows just a sample, it means when you submit a notebook the hidden set will replace it.So you should take care about the full set. Here is the description in the data page:\n>This competition uses a hidden test. When your submitted notebook is scored, the actual test data (including a full length sample submission) will be made available to your notebook.",
    "2846653": "Thanks, useful comment and source of my problems!"
  },
  "source": "meta"
}