{
  "id": 492712,
  "title": "CountEncoding and Overfitting",
  "url": "/competitions/home-credit-credit-risk-model-stability/discussion/492712",
  "author_name": "Denis Platonov UA",
  "post_date": "2024-04-10T17:06:50.997000",
  "votes": 9,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I'm curious about recent CountEncoding notebooks that have achieved a public score of 0.579. Are these models potentially overfitting to the public leaderboard, or is there a reason they are performing so well?</p>",
  "messages": [
    {
      "id": 2745522,
      "postDate": "2024-04-10T17:06:50.997Z",
      "content": "<p>I'm curious about recent CountEncoding notebooks that have achieved a public score of 0.579. Are these models potentially overfitting to the public leaderboard, or is there a reason they are performing so well?</p>",
      "rawMarkdown": "I'm curious about recent CountEncoding notebooks that have achieved a public score of 0.579. Are these models potentially overfitting to the public leaderboard, or is there a reason they are performing so well?",
      "votes": 9
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2745522": "I'm curious about recent CountEncoding notebooks that have achieved a public score of 0.579. Are these models potentially overfitting to the public leaderboard, or is there a reason they are performing so well?"
  }
}