{
  "id": 582696,
  "title": "Questions about subjects and acceleration",
  "url": "/competitions/cmi-detect-behavior-with-sensor-data/discussion/582696",
  "author_name": "",
  "post_date": "2025-06-02T09:11:22.402571200Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>If I understand correctly, this is not a study on real subjects who have these BFRBs, but rather with people who mimic this behavior. The question is whether the results will not be affected by this. </p>\n<p>And I'm not sure about this, but the acceleration values seem quite high while the described positions of the subjects are seated, lied on back, etc. While is expected to record some outliers, or maybe have gravity included, about half of the records have total acceleration over 10 m/s2 and they are not only on z axis. Which is your perspective about this? </p>",
  "messages": [
    {
      "id": "3215457",
      "postDate": "06/02/2025 09:11:22",
      "content": "<p>If I understand correctly, this is not a study on real subjects who have these BFRBs, but rather with people who mimic this behavior. The question is whether the results will not be affected by this. </p>\n<p>And I'm not sure about this, but the acceleration values seem quite high while the described positions of the subjects are seated, lied on back, etc. While is expected to record some outliers, or maybe have gravity included, about half of the records have total acceleration over 10 m/s2 and they are not only on z axis. Which is your perspective about this? </p>",
      "rawMarkdown": "If I understand correctly, this is not a study on real subjects who have these BFRBs, but rather with people who mimic this behavior. The question is whether the results will not be affected by this. \n\nAnd I'm not sure about this, but the acceleration values seem quite high while the described positions of the subjects are seated, lied on back, etc. While is expected to record some outliers, or maybe have gravity included, about half of the records have total acceleration over 10 m/s2 and they are not only on z axis. Which is your perspective about this?",
      "votes": null
    },
    {
      "id": "3215558",
      "postDate": "06/02/2025 12:13:41",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/cristinadanielapopa\" target=\"_blank\">@cristinadanielapopa</a>, that's a good question but looking at the values, I concluded that the gravity is included in the sensor measurments. For a few samples that I checked, the length of the acceleration vector is close to 9.81 which equals g. The reason why this isn't all in the z direction is because the watch isn't places horizontally so it's reference frame is rotated compared to the earth surface.</p>",
      "rawMarkdown": "Hi @cristinadanielapopa, that's a good question but looking at the values, I concluded that the gravity is included in the sensor measurments. For a few samples that I checked, the length of the acceleration vector is close to 9.81 which equals g. The reason why this isn't all in the z direction is because the watch isn't places horizontally so it's reference frame is rotated compared to the earth surface.",
      "votes": null
    },
    {
      "id": "3215567",
      "postDate": "06/02/2025 12:30:24",
      "content": "<p>Thank you <a href=\"https://www.kaggle.com/tom3141592\" target=\"_blank\">@tom3141592</a>! Make sense.</p>",
      "rawMarkdown": "Thank you @tom3141592! Make sense.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3215558,
      "author_name": "tom3141592",
      "author_url": "",
      "post_date": "06/02/2025 12:13:41",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/cristinadanielapopa\" target=\"_blank\">@cristinadanielapopa</a>, that's a good question but looking at the values, I concluded that the gravity is included in the sensor measurments. For a few samples that I checked, the length of the acceleration vector is close to 9.81 which equals g. The reason why this isn't all in the z direction is because the watch isn't places horizontally so it's reference frame is rotated compared to the earth surface.</p>",
      "votes": null,
      "replies": [
        {
          "id": 3215567,
          "author_name": "cristinadanielapopa",
          "author_url": "",
          "post_date": "06/02/2025 12:30:24",
          "content": "<p>Thank you <a href=\"https://www.kaggle.com/tom3141592\" target=\"_blank\">@tom3141592</a>! Make sense.</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "3215457": "If I understand correctly, this is not a study on real subjects who have these BFRBs, but rather with people who mimic this behavior. The question is whether the results will not be affected by this. \n\nAnd I'm not sure about this, but the acceleration values seem quite high while the described positions of the subjects are seated, lied on back, etc. While is expected to record some outliers, or maybe have gravity included, about half of the records have total acceleration over 10 m/s2 and they are not only on z axis. Which is your perspective about this?",
    "3215558": "Hi @cristinadanielapopa, that's a good question but looking at the values, I concluded that the gravity is included in the sensor measurments. For a few samples that I checked, the length of the acceleration vector is close to 9.81 which equals g. The reason why this isn't all in the z direction is because the watch isn't places horizontally so it's reference frame is rotated compared to the earth surface.",
    "3215567": "Thank you @tom3141592! Make sense."
  },
  "source": "meta"
}