{
  "id": 548934,
  "title": "Any inputs on how the smartwatch features like hr, steps, carbs etc improve the prediction results?",
  "url": "/competitions/brist1d/discussion/548934",
  "author_name": "",
  "post_date": "2024-11-29T15:32:48.094521800Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I just wanted to see if anyone has probed more into how much impact these smartwatch features provide on the prediction compared to historical blood glucose and insulin data. I believe the main goal of this competition is to see how better of a model we end up with compared to the traditional setup that uses only historical blood glucose data.</p>",
  "messages": [
    {
      "id": "3058555",
      "postDate": "11/29/2024 15:32:48",
      "content": "<p>I just wanted to see if anyone has probed more into how much impact these smartwatch features provide on the prediction compared to historical blood glucose and insulin data. I believe the main goal of this competition is to see how better of a model we end up with compared to the traditional setup that uses only historical blood glucose data.</p>",
      "rawMarkdown": "I just wanted to see if anyone has probed more into how much impact these smartwatch features provide on the prediction compared to historical blood glucose and insulin data. I believe the main goal of this competition is to see how better of a model we end up with compared to the traditional setup that uses only historical blood glucose data.",
      "votes": null
    },
    {
      "id": "3058692",
      "postDate": "11/29/2024 19:46:17",
      "content": "<p>After trying various configurations and feature engineering, I can most certainly say that the smartwatch features only play a minor role (around 0.1 RMSE). Sometimes I can even get better accuracy dropping columns like those related to carbs.</p>",
      "rawMarkdown": "After trying various configurations and feature engineering, I can most certainly say that the smartwatch features only play a minor role (around 0.1 RMSE). Sometimes I can even get better accuracy dropping columns like those related to carbs.",
      "votes": null
    },
    {
      "id": "3058795",
      "postDate": "11/30/2024 01:51:15",
      "content": "<p>I am also getting similar results only.</p>",
      "rawMarkdown": "I am also getting similar results only.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3058692,
      "author_name": "alrjandrorivasespa",
      "author_url": "",
      "post_date": "11/29/2024 19:46:17",
      "content": "<p>After trying various configurations and feature engineering, I can most certainly say that the smartwatch features only play a minor role (around 0.1 RMSE). Sometimes I can even get better accuracy dropping columns like those related to carbs.</p>",
      "votes": null,
      "replies": [
        {
          "id": 3058795,
          "author_name": "msundarv",
          "author_url": "",
          "post_date": "11/30/2024 01:51:15",
          "content": "<p>I am also getting similar results only.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "3058555": "I just wanted to see if anyone has probed more into how much impact these smartwatch features provide on the prediction compared to historical blood glucose and insulin data. I believe the main goal of this competition is to see how better of a model we end up with compared to the traditional setup that uses only historical blood glucose data.",
    "3058692": "After trying various configurations and feature engineering, I can most certainly say that the smartwatch features only play a minor role (around 0.1 RMSE). Sometimes I can even get better accuracy dropping columns like those related to carbs.",
    "3058795": "I am also getting similar results only."
  },
  "source": "meta"
}