{
  "id": 466178,
  "title": "What is the Evaluation Weight? ",
  "url": "/competitions/aocr2024/discussion/466178",
  "author_name": "",
  "post_date": "2024-01-07T13:29:07.399971900Z",
  "votes": null,
  "comment_count": 1,
  "views": 0,
  "content": "<blockquote>\n  <p>Each CT slice is evaluated to identify the presence of an abnormal appendix indicative of appendicitis. Furthermore, the entire CT scan is assessed to determine a diagnosis of acute appendicitis for that particular scan session.</p>\n</blockquote>\n<p>Submissions will be assessed based on the F1 score, comparing the predictions to the actual targets.</p>\n<p>Two distinct F1 scores will be determined: one for the scan-level predictions and another for the slice-level predictions. The final leaderboard ranking will be derived from a composite of these two F1 scores.</p>\n<p>Hello, what is the weight value of scan-level predictions and slice-level prediction? <br>\nThere is 200 user and 19153 - 200 slice in sample_submission. <br>\nThen, </p>\n<ul>\n<li>0.5 * F1(scan - 200 rows) + 0.5 * F1(slice  - 18953 rows) <br>\nor </li>\n<li>1 * F1(scan + slice - 19153 rows) ?</li>\n</ul>\n<p>Thanks</p>",
  "messages": [
    {
      "id": "2590843",
      "postDate": "01/07/2024 13:29:07",
      "content": "<blockquote>\n  <p>Each CT slice is evaluated to identify the presence of an abnormal appendix indicative of appendicitis. Furthermore, the entire CT scan is assessed to determine a diagnosis of acute appendicitis for that particular scan session.</p>\n</blockquote>\n<p>Submissions will be assessed based on the F1 score, comparing the predictions to the actual targets.</p>\n<p>Two distinct F1 scores will be determined: one for the scan-level predictions and another for the slice-level predictions. The final leaderboard ranking will be derived from a composite of these two F1 scores.</p>\n<p>Hello, what is the weight value of scan-level predictions and slice-level prediction? <br>\nThere is 200 user and 19153 - 200 slice in sample_submission. <br>\nThen, </p>\n<ul>\n<li>0.5 * F1(scan - 200 rows) + 0.5 * F1(slice  - 18953 rows) <br>\nor </li>\n<li>1 * F1(scan + slice - 19153 rows) ?</li>\n</ul>\n<p>Thanks</p>",
      "rawMarkdown": ">Each CT slice is evaluated to identify the presence of an abnormal appendix indicative of appendicitis. Furthermore, the entire CT scan is assessed to determine a diagnosis of acute appendicitis for that particular scan session.\n\nSubmissions will be assessed based on the F1 score, comparing the predictions to the actual targets.\n\nTwo distinct F1 scores will be determined: one for the scan-level predictions and another for the slice-level predictions. The final leaderboard ranking will be derived from a composite of these two F1 scores.\n\nHello, what is the weight value of scan-level predictions and slice-level prediction? \nThere is 200 user and 19153 - 200 slice in sample_submission. \nThen, \n- 0.5 * F1(scan - 200 rows) + 0.5 * F1(slice  - 18953 rows) \nor \n- 1 * F1(scan + slice - 19153 rows) ?\n\nThanks",
      "votes": null
    },
    {
      "id": "2591446",
      "postDate": "01/08/2024 01:03:40",
      "content": "<p>final score = (scan-level F1) *0.5 + (slice-level F1) *0.5</p>",
      "rawMarkdown": "final score = (scan-level F1) *0.5 + (slice-level F1) *0.5",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2591446,
      "author_name": "goman178",
      "author_url": "",
      "post_date": "01/08/2024 01:03:40",
      "content": "<p>final score = (scan-level F1) *0.5 + (slice-level F1) *0.5</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2590843": ">Each CT slice is evaluated to identify the presence of an abnormal appendix indicative of appendicitis. Furthermore, the entire CT scan is assessed to determine a diagnosis of acute appendicitis for that particular scan session.\n\nSubmissions will be assessed based on the F1 score, comparing the predictions to the actual targets.\n\nTwo distinct F1 scores will be determined: one for the scan-level predictions and another for the slice-level predictions. The final leaderboard ranking will be derived from a composite of these two F1 scores.\n\nHello, what is the weight value of scan-level predictions and slice-level prediction? \nThere is 200 user and 19153 - 200 slice in sample_submission. \nThen, \n- 0.5 * F1(scan - 200 rows) + 0.5 * F1(slice  - 18953 rows) \nor \n- 1 * F1(scan + slice - 19153 rows) ?\n\nThanks",
    "2591446": "final score = (scan-level F1) *0.5 + (slice-level F1) *0.5"
  },
  "source": "meta"
}