{
  "id": 221826,
  "title": "Why are CVC classes hard to predict?",
  "url": "/competitions/ranzcr-clip-catheter-line-classification/discussion/221826",
  "author_name": "",
  "post_date": "2021-02-24T06:45:05.261184500Z",
  "votes": 6,
  "comment_count": 2,
  "views": 0,
  "content": "<p>As an example, here is a model I am training and so far, here is the cross-validation scores I get for the 11 classes for one fold: </p>\n<pre><code>[0.9412 0.9308 0.9849 0.8808 0.8782 0.9633 0.9586 0.628  0.6372 0.6281\n 0.9871]\n</code></pre>\n<p>All classes have pretty high scores except <strong>CVC</strong> (the three before the last one).</p>\n<p>Is there an explanation why this happens? Since I am new to this challenge, some insights have probably been shared, if anyone can direct me to the correct discussions, that would be very appreciated. Thanks! :)</p>",
  "messages": [
    {
      "id": "1216229",
      "postDate": "02/24/2021 06:45:05",
      "content": "<p>As an example, here is a model I am training and so far, here is the cross-validation scores I get for the 11 classes for one fold: </p>\n<pre><code>[0.9412 0.9308 0.9849 0.8808 0.8782 0.9633 0.9586 0.628  0.6372 0.6281\n 0.9871]\n</code></pre>\n<p>All classes have pretty high scores except <strong>CVC</strong> (the three before the last one).</p>\n<p>Is there an explanation why this happens? Since I am new to this challenge, some insights have probably been shared, if anyone can direct me to the correct discussions, that would be very appreciated. Thanks! :)</p>",
      "rawMarkdown": "As an example, here is a model I am training and so far, here is the cross-validation scores I get for the 11 classes for one fold: \n\n```\n[0.9412 0.9308 0.9849 0.8808 0.8782 0.9633 0.9586 0.628  0.6372 0.6281\n 0.9871]\n```\n\nAll classes have pretty high scores except **CVC** (the three before the last one).\n\nIs there an explanation why this happens? Since I am new to this challenge, some insights have probably been shared, if anyone can direct me to the correct discussions, that would be very appreciated. Thanks! :)",
      "votes": null
    },
    {
      "id": "1216243",
      "postDate": "02/24/2021 06:49:24",
      "content": "<p>Hello!</p>\n<p>There's actually a <strong><a href=\"https://www.kaggle.com/c/ranzcr-clip-catheter-line-classification/discussion/216219\" target=\"_blank\">great explanation</a></strong> by <a href=\"https://www.kaggle.com/ryches\" target=\"_blank\">@ryches</a> on why CVC is actually your <em>best</em> performing classifier.</p>",
      "rawMarkdown": "Hello!\n\nThere's actually a **[great explanation](https://www.kaggle.com/c/ranzcr-clip-catheter-line-classification/discussion/216219)** by @ryches on why CVC is actually your *best* performing classifier.",
      "votes": null
    },
    {
      "id": "1216249",
      "postDate": "02/24/2021 06:54:11",
      "content": "<p>Oh, thanks very much for the link. 👀</p>",
      "rawMarkdown": "Oh, thanks very much for the link. 👀",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1216243,
      "author_name": "nickuzmenkov",
      "author_url": "",
      "post_date": "02/24/2021 06:49:24",
      "content": "<p>Hello!</p>\n<p>There's actually a <strong><a href=\"https://www.kaggle.com/c/ranzcr-clip-catheter-line-classification/discussion/216219\" target=\"_blank\">great explanation</a></strong> by <a href=\"https://www.kaggle.com/ryches\" target=\"_blank\">@ryches</a> on why CVC is actually your <em>best</em> performing classifier.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1216249,
          "author_name": "yassinealouini",
          "author_url": "",
          "post_date": "02/24/2021 06:54:11",
          "content": "<p>Oh, thanks very much for the link. 👀</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1216229": "As an example, here is a model I am training and so far, here is the cross-validation scores I get for the 11 classes for one fold: \n\n```\n[0.9412 0.9308 0.9849 0.8808 0.8782 0.9633 0.9586 0.628  0.6372 0.6281\n 0.9871]\n```\n\nAll classes have pretty high scores except **CVC** (the three before the last one).\n\nIs there an explanation why this happens? Since I am new to this challenge, some insights have probably been shared, if anyone can direct me to the correct discussions, that would be very appreciated. Thanks! :)",
    "1216243": "Hello!\n\nThere's actually a **[great explanation](https://www.kaggle.com/c/ranzcr-clip-catheter-line-classification/discussion/216219)** by @ryches on why CVC is actually your *best* performing classifier.",
    "1216249": "Oh, thanks very much for the link. 👀"
  },
  "source": "meta"
}