{
  "id": 394003,
  "title": "Strongest predictors",
  "url": "/competitions/tlvmc-parkinsons-freezing-gait-prediction/discussion/394003",
  "author_name": "FelipeKitamura, MD, PhD",
  "post_date": "2023-03-11T16:43:04.031000",
  "votes": 1,
  "comment_count": 0,
  "views": 0,
  "content": "<p>After reading the documentation of this competition, my opinion as a physician is that the strongest predictors in the entire dataset should be the accelerometer time series (X, Y, Z) as explained in <a href=\"https://www.kaggle.com/competitions/tlvmc-parkinsons-freezing-gait-prediction/discussion/393852\" target=\"_blank\">this post</a></p>\n<p>I am not saying they are the only predictors. The other clinical variables might also have some signal, but I think a good baseline model based only on the accelerometer information should do well enough in this task. However, doing well enough might be insufficient to win this competition.</p>",
  "messages": [
    {
      "id": 2177648,
      "postDate": "2023-03-11T16:43:04.030Z",
      "content": "<p>After reading the documentation of this competition, my opinion as a physician is that the strongest predictors in the entire dataset should be the accelerometer time series (X, Y, Z) as explained in <a href=\"https://www.kaggle.com/competitions/tlvmc-parkinsons-freezing-gait-prediction/discussion/393852\" target=\"_blank\">this post</a></p>\n<p>I am not saying they are the only predictors. The other clinical variables might also have some signal, but I think a good baseline model based only on the accelerometer information should do well enough in this task. However, doing well enough might be insufficient to win this competition.</p>",
      "rawMarkdown": "After reading the documentation of this competition, my opinion as a physician is that the strongest predictors in the entire dataset should be the accelerometer time series (X, Y, Z) as explained in [this post](https://www.kaggle.com/competitions/tlvmc-parkinsons-freezing-gait-prediction/discussion/393852)\n\nI am not saying they are the only predictors. The other clinical variables might also have some signal, but I think a good baseline model based only on the accelerometer information should do well enough in this task. However, doing well enough might be insufficient to win this competition.",
      "votes": 1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2177648": "After reading the documentation of this competition, my opinion as a physician is that the strongest predictors in the entire dataset should be the accelerometer time series (X, Y, Z) as explained in [this post](https://www.kaggle.com/competitions/tlvmc-parkinsons-freezing-gait-prediction/discussion/393852)\n\nI am not saying they are the only predictors. The other clinical variables might also have some signal, but I think a good baseline model based only on the accelerometer information should do well enough in this task. However, doing well enough might be insufficient to win this competition."
  }
}