{
  "id": 145619,
  "title": "PDF of PANDA challenge - structured description",
  "url": "/competitions/prostate-cancer-grade-assessment/discussion/145619",
  "author_name": "Rafi Hai",
  "post_date": "2020-04-23T21:39:07.728000",
  "votes": 3,
  "comment_count": 1,
  "views": 0,
  "content": "<p>The hosts shared an important <a href=\"https://zenodo.org/record/3715938#.XqIJEWgzZPb\">link </a>which apparently is only located in the Data section. \nIt describes among others the total number of training, validation and test cases:\n<code>\nTraining set: +/- 11.000 cases\nPublic test set: +/- 400 cases (with expert gradings)\nPrivate test set: +/- 400 cases (with expert gradings)\n</code>\nYou can find more detailed info about the annotation characteristics of Radboudumc vs. Karolinska, the sources of error etc. Definitely worth to check out.</p>",
  "messages": [
    {
      "id": 818421,
      "postDate": "2020-04-23T21:39:07.730Z",
      "content": "<p>The hosts shared an important <a href=\"https://zenodo.org/record/3715938#.XqIJEWgzZPb\">link </a>which apparently is only located in the Data section. \nIt describes among others the total number of training, validation and test cases:\n<code>\nTraining set: +/- 11.000 cases\nPublic test set: +/- 400 cases (with expert gradings)\nPrivate test set: +/- 400 cases (with expert gradings)\n</code>\nYou can find more detailed info about the annotation characteristics of Radboudumc vs. Karolinska, the sources of error etc. Definitely worth to check out.</p>",
      "rawMarkdown": "The hosts shared an important [link ](https://zenodo.org/record/3715938#.XqIJEWgzZPb)which apparently is only located in the Data section. \nIt describes among others the total number of training, validation and test cases:\n```\nTraining set: +/- 11.000 cases\nPublic test set: +/- 400 cases (with expert gradings)\nPrivate test set: +/- 400 cases (with expert gradings)\n```\nYou can find more detailed info about the annotation characteristics of Radboudumc vs. Karolinska, the sources of error etc. Definitely worth to check out.",
      "votes": 3
    },
    {
      "id": 818423,
      "postDate": "2020-04-23T21:43:20.137Z",
      "content": "<p>thanks :)</p>",
      "rawMarkdown": "thanks :)"
    }
  ],
  "comments": [
    {
      "id": 818423,
      "author_name": "Alberto Maria Falletta",
      "author_url": "",
      "post_date": "2020-04-23T21:43:20.137000",
      "content": "<p>thanks :)</p>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "818421": "The hosts shared an important [link ](https://zenodo.org/record/3715938#.XqIJEWgzZPb)which apparently is only located in the Data section. \nIt describes among others the total number of training, validation and test cases:\n```\nTraining set: +/- 11.000 cases\nPublic test set: +/- 400 cases (with expert gradings)\nPrivate test set: +/- 400 cases (with expert gradings)\n```\nYou can find more detailed info about the annotation characteristics of Radboudumc vs. Karolinska, the sources of error etc. Definitely worth to check out.",
    "818423": "thanks :)"
  }
}