{
  "id": 169132,
  "title": "Any explanation for the shake up?",
  "url": "/competitions/prostate-cancer-grade-assessment/discussion/169132",
  "author_name": "FelipeKitamura, MD, PhD",
  "post_date": "2020-07-23T02:30:05.776000",
  "votes": 2,
  "comment_count": 3,
  "views": 0,
  "content": "<p>I can't come up with a reasonable explanation for this impressive shakeup. any thoughts?</p>",
  "messages": [
    {
      "id": 940562,
      "postDate": "2020-07-23T02:56:34.390Z",
      "content": "<ol>\n<li>Allegedly, the train and test set labeling standards are different</li>\n<li>Test set is too small, only 940 (public and private combined) images</li>\n<li>qwk metric is very sensitive to thresholds</li>\n<li>We can't see the test images. So who knows what's in there </li>\n</ol>",
      "rawMarkdown": "1. Allegedly, the train and test set labeling standards are different\n2. Test set is too small, only 940 (public and private combined) images\n3. qwk metric is very sensitive to thresholds\n4. We can't see the test images. So who knows what's in there ",
      "votes": 1
    },
    {
      "id": 940539,
      "postDate": "2020-07-23T02:30:05.777Z",
      "content": "<p>I can't come up with a reasonable explanation for this impressive shakeup. any thoughts?</p>",
      "rawMarkdown": "I can't come up with a reasonable explanation for this impressive shakeup. any thoughts?",
      "votes": 2
    },
    {
      "id": 941619,
      "postDate": "2020-07-23T10:26:50.910Z",
      "content": "<p>It may be because of the way the test data was split into public and private. There is likely a significant difference in the data distribution. It would be nice to hear more on this from the organizers</p>",
      "rawMarkdown": "It may be because of the way the test data was split into public and private. There is likely a significant difference in the data distribution. It would be nice to hear more on this from the organizers"
    },
    {
      "id": 941526,
      "postDate": "2020-07-23T09:16:56.367Z",
      "content": "<ul>\n<li>Issue of noise , those who worked got better and money.</li>\n<li>Complex model scored less and simple model did better job.</li>\n</ul>",
      "rawMarkdown": "- Issue of noise , those who worked got better and money.\n- Complex model scored less and simple model did better job."
    }
  ],
  "comments": [
    {
      "id": 940562,
      "author_name": "Bo",
      "author_url": "",
      "post_date": "2020-07-23T02:56:34.390000",
      "content": "<ol>\n<li>Allegedly, the train and test set labeling standards are different</li>\n<li>Test set is too small, only 940 (public and private combined) images</li>\n<li>qwk metric is very sensitive to thresholds</li>\n<li>We can't see the test images. So who knows what's in there </li>\n</ol>",
      "votes": 1,
      "replies": []
    },
    {
      "id": 941619,
      "author_name": "Pasquale",
      "author_url": "",
      "post_date": "2020-07-23T10:26:50.910000",
      "content": "<p>It may be because of the way the test data was split into public and private. There is likely a significant difference in the data distribution. It would be nice to hear more on this from the organizers</p>",
      "votes": 0,
      "replies": []
    },
    {
      "id": 941526,
      "author_name": "Rajnish Chauhan",
      "author_url": "",
      "post_date": "2020-07-23T09:16:56.367000",
      "content": "<ul>\n<li>Issue of noise , those who worked got better and money.</li>\n<li>Complex model scored less and simple model did better job.</li>\n</ul>",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "940562": "1. Allegedly, the train and test set labeling standards are different\n2. Test set is too small, only 940 (public and private combined) images\n3. qwk metric is very sensitive to thresholds\n4. We can't see the test images. So who knows what's in there ",
    "940539": "I can't come up with a reasonable explanation for this impressive shakeup. any thoughts?",
    "941619": "It may be because of the way the test data was split into public and private. There is likely a significant difference in the data distribution. It would be nice to hear more on this from the organizers",
    "941526": "- Issue of noise , those who worked got better and money.\n- Complex model scored less and simple model did better job."
  }
}