{
  "id": 295024,
  "title": "Models performance differ a lot by the presence of tag data",
  "url": "/competitions/cs5785-fall-2021-final/discussion/295024",
  "author_name": "Gavin Hu",
  "post_date": "2021-12-14T02:23:31.284000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>During our experiment, we found that each time when we add tags data to the model using feature matrix only (with other methods remaining constant), the Kaggle score improves from 0.2x to almost 0.4. So tags data seems to be crucial. Hope this helps.</p>",
  "messages": [
    {
      "id": 1617278,
      "postDate": "2021-12-14T02:23:31.283Z",
      "content": "<p>During our experiment, we found that each time when we add tags data to the model using feature matrix only (with other methods remaining constant), the Kaggle score improves from 0.2x to almost 0.4. So tags data seems to be crucial. Hope this helps.</p>",
      "rawMarkdown": "During our experiment, we found that each time when we add tags data to the model using feature matrix only (with other methods remaining constant), the Kaggle score improves from 0.2x to almost 0.4. So tags data seems to be crucial. Hope this helps."
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "1617278": "During our experiment, we found that each time when we add tags data to the model using feature matrix only (with other methods remaining constant), the Kaggle score improves from 0.2x to almost 0.4. So tags data seems to be crucial. Hope this helps."
  }
}