{
  "id": 349825,
  "title": "Clarify hidden test set size - is really 1500 images?",
  "url": "/competitions/rsna-2022-cervical-spine-fracture-detection/discussion/349825",
  "author_name": "SolverWorld",
  "post_date": "2022-09-02T22:12:01.091000",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>In the Data section<br>\n<a href=\"https://www.kaggle.com/competitions/rsna-2022-cervical-spine-fracture-detection/data\" target=\"_blank\">https://www.kaggle.com/competitions/rsna-2022-cervical-spine-fracture-detection/data</a><br>\nit says to expect 1500 scans in the hidden test set.</p>\n<p>Is that 1500 images (dcm) files?  If so, and with roughly 300 scans per patient, that means the test set has only 5 patients in it.  The leaderboard page says that the public LB is calculated with 28% of the data, making it calculated on about 1.4 patients! </p>\n<p>That would make for a very noisy leaderboard.</p>\n<p>Or does it mean 1500 patients, which would mean around 450,000 images?</p>\n<p>Or maybe they mean the test.csv file will have 1500 entries, which would be 1500/7 = 214 patients, so maybe 214*300=64,200 images.</p>",
  "messages": [
    {
      "id": 1924283,
      "postDate": "2022-09-02T22:12:01.093Z",
      "content": "<p>In the Data section<br>\n<a href=\"https://www.kaggle.com/competitions/rsna-2022-cervical-spine-fracture-detection/data\" target=\"_blank\">https://www.kaggle.com/competitions/rsna-2022-cervical-spine-fracture-detection/data</a><br>\nit says to expect 1500 scans in the hidden test set.</p>\n<p>Is that 1500 images (dcm) files?  If so, and with roughly 300 scans per patient, that means the test set has only 5 patients in it.  The leaderboard page says that the public LB is calculated with 28% of the data, making it calculated on about 1.4 patients! </p>\n<p>That would make for a very noisy leaderboard.</p>\n<p>Or does it mean 1500 patients, which would mean around 450,000 images?</p>\n<p>Or maybe they mean the test.csv file will have 1500 entries, which would be 1500/7 = 214 patients, so maybe 214*300=64,200 images.</p>",
      "rawMarkdown": "In the Data section\nhttps://www.kaggle.com/competitions/rsna-2022-cervical-spine-fracture-detection/data\nit says to expect 1500 scans in the hidden test set.\n\nIs that 1500 images (dcm) files?  If so, and with roughly 300 scans per patient, that means the test set has only 5 patients in it.  The leaderboard page says that the public LB is calculated with 28% of the data, making it calculated on about 1.4 patients! \n\nThat would make for a very noisy leaderboard.\n\nOr does it mean 1500 patients, which would mean around 450,000 images?\n\nOr maybe they mean the test.csv file will have 1500 entries, which would be 1500/7 = 214 patients, so maybe 214*300=64,200 images.",
      "votes": 1
    },
    {
      "id": 1924284,
      "postDate": "2022-09-02T22:18:42.893Z",
      "content": "<p>It is 1500 patients…</p>",
      "rawMarkdown": "It is 1500 patients...",
      "votes": 2,
      "replies": [
        {
          "id": 1924287,
          "postDate": "2022-09-02T22:30:13.243Z",
          "content": "<p>Thanks, that makes a lot more sense</p>",
          "rawMarkdown": "Thanks, that makes a lot more sense"
        }
      ]
    }
  ],
  "comments": [
    {
      "id": 1924284,
      "author_name": "Harshit Sheoran",
      "author_url": "",
      "post_date": "2022-09-02T22:18:42.893000",
      "content": "<p>It is 1500 patients…</p>",
      "votes": 2,
      "replies": [
        {
          "id": 1924287,
          "author_name": "SolverWorld",
          "author_url": "",
          "post_date": "2022-09-02T22:30:13.243000",
          "content": "<p>Thanks, that makes a lot more sense</p>",
          "votes": 0,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1924283": "In the Data section\nhttps://www.kaggle.com/competitions/rsna-2022-cervical-spine-fracture-detection/data\nit says to expect 1500 scans in the hidden test set.\n\nIs that 1500 images (dcm) files?  If so, and with roughly 300 scans per patient, that means the test set has only 5 patients in it.  The leaderboard page says that the public LB is calculated with 28% of the data, making it calculated on about 1.4 patients! \n\nThat would make for a very noisy leaderboard.\n\nOr does it mean 1500 patients, which would mean around 450,000 images?\n\nOr maybe they mean the test.csv file will have 1500 entries, which would be 1500/7 = 214 patients, so maybe 214*300=64,200 images.",
    "1924284": "It is 1500 patients..."
  }
}