{
  "id": 145278,
  "title": "Optimizer for QWK Metric",
  "url": "/competitions/prostate-cancer-grade-assessment/discussion/145278",
  "author_name": "Ian Pan",
  "post_date": "2020-04-22T14:24:32.160000",
  "votes": 10,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I found this great kernel by @abhishek from the APTOS competition:</p>\n\n<p><a href=\"https://www.kaggle.com/abhishek/optimizer-for-quadratic-weighted-kappa\">https://www.kaggle.com/abhishek/optimizer-for-quadratic-weighted-kappa</a></p>\n\n<p>If you are using a regression approach, this allows you to optimize thresholds for assigning classes by tuning on a validation set. Be careful of overfitting.</p>",
  "messages": [
    {
      "id": 816707,
      "postDate": "2020-04-22T14:24:32.160Z",
      "content": "<p>I found this great kernel by @abhishek from the APTOS competition:</p>\n\n<p><a href=\"https://www.kaggle.com/abhishek/optimizer-for-quadratic-weighted-kappa\">https://www.kaggle.com/abhishek/optimizer-for-quadratic-weighted-kappa</a></p>\n\n<p>If you are using a regression approach, this allows you to optimize thresholds for assigning classes by tuning on a validation set. Be careful of overfitting.</p>",
      "rawMarkdown": "I found this great kernel by @abhishek from the APTOS competition:\n\nhttps://www.kaggle.com/abhishek/optimizer-for-quadratic-weighted-kappa\n\nIf you are using a regression approach, this allows you to optimize thresholds for assigning classes by tuning on a validation set. Be careful of overfitting.",
      "votes": 10
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "816707": "I found this great kernel by @abhishek from the APTOS competition:\n\nhttps://www.kaggle.com/abhishek/optimizer-for-quadratic-weighted-kappa\n\nIf you are using a regression approach, this allows you to optimize thresholds for assigning classes by tuning on a validation set. Be careful of overfitting."
  }
}