{
  "id": 249190,
  "title": "What data are you using?",
  "url": "/competitions/google-smartphone-decimeter-challenge/discussion/249190",
  "author_name": "",
  "post_date": "2021-06-27T05:28:10.855683900Z",
  "votes": 3,
  "comment_count": 9,
  "views": 0,
  "content": "<p>So far I used ground truth datasets with the baseline train/test only, I could not (yet) improve my score with any other file like the derived or raw gnss data.</p>\n<p>What datasets did you find beneficial?</p>",
  "messages": [
    {
      "id": "1366691",
      "postDate": "06/27/2021 05:28:10",
      "content": "<p>So far I used ground truth datasets with the baseline train/test only, I could not (yet) improve my score with any other file like the derived or raw gnss data.</p>\n<p>What datasets did you find beneficial?</p>",
      "rawMarkdown": "So far I used ground truth datasets with the baseline train/test only, I could not (yet) improve my score with any other file like the derived or raw gnss data.\n\nWhat datasets did you find beneficial?",
      "votes": null
    },
    {
      "id": "1367557",
      "postDate": "06/27/2021 20:50:03",
      "content": "<p>Just used everything and threw it into a scaler and an automl tool. Idc much about this comp.</p>",
      "rawMarkdown": "Just used everything and threw it into a scaler and an automl tool. Idc much about this comp.",
      "votes": null
    },
    {
      "id": "1367631",
      "postDate": "06/28/2021 01:05:10",
      "content": "<p>For LB score, I only using baseline.  <br>\nAppropriate treatment of the outliers that already discussed on discussions or notebook can improve accuracy by only baseline data.</p>",
      "rawMarkdown": "For LB score, I only using baseline.  \nAppropriate treatment of the outliers that already discussed on discussions or notebook can improve accuracy by only baseline data.",
      "votes": null
    },
    {
      "id": "1367704",
      "postDate": "06/28/2021 03:35:01",
      "content": "<p>Wow thats amazing for a LB like this, mabey this is the best approach if so..</p>",
      "rawMarkdown": "Wow thats amazing for a LB like this, mabey this is the best approach if so..",
      "votes": null
    },
    {
      "id": "1367705",
      "postDate": "06/28/2021 03:35:51",
      "content": "<p>Thanks!<br>\nIs auto ML legit in this competition?</p>",
      "rawMarkdown": "Thanks!\nIs auto ML legit in this competition?",
      "votes": null
    },
    {
      "id": "1368104",
      "postDate": "06/28/2021 10:29:10",
      "content": "<p>gnss log has data about the movement of the vehicle. This is making my model better.</p>",
      "rawMarkdown": "gnss log has data about the movement of the vehicle. This is making my model better.",
      "votes": null
    },
    {
      "id": "1368146",
      "postDate": "06/28/2021 11:03:22",
      "content": "<p>awsome thank you!</p>",
      "rawMarkdown": "awsome thank you!",
      "votes": null
    },
    {
      "id": "1368375",
      "postDate": "06/28/2021 14:32:56",
      "content": "<p>Nothing but a baseline, because I still think that I havent used the baseline to the full extent yet. <br>\nI am planning to incorporate the gnss logs by the end of the week though.</p>",
      "rawMarkdown": "Nothing but a baseline, because I still think that I havent used the baseline to the full extent yet. \nI am planning to incorporate the gnss logs by the end of the week though.",
      "votes": null
    },
    {
      "id": "1368468",
      "postDate": "06/28/2021 15:43:18",
      "content": "<p>That's very impressive, thank you for sharing!</p>",
      "rawMarkdown": "That's very impressive, thank you for sharing!",
      "votes": null
    },
    {
      "id": "1369886",
      "postDate": "06/29/2021 17:17:54",
      "content": "<p>Not really</p>",
      "rawMarkdown": "Not really",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1367557,
      "author_name": "eladwar",
      "author_url": "",
      "post_date": "06/27/2021 20:50:03",
      "content": "<p>Just used everything and threw it into a scaler and an automl tool. Idc much about this comp.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1367705,
          "author_name": "avivlevi815",
          "author_url": "",
          "post_date": "06/28/2021 03:35:51",
          "content": "<p>Thanks!<br>\nIs auto ML legit in this competition?</p>",
          "votes": null,
          "replies": []
        },
        {
          "id": 1369886,
          "author_name": "avtobusbratiev",
          "author_url": "",
          "post_date": "06/29/2021 17:17:54",
          "content": "<p>Not really</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1367631,
      "author_name": "mrt0933",
      "author_url": "",
      "post_date": "06/28/2021 01:05:10",
      "content": "<p>For LB score, I only using baseline.  <br>\nAppropriate treatment of the outliers that already discussed on discussions or notebook can improve accuracy by only baseline data.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1367704,
          "author_name": "avivlevi815",
          "author_url": "",
          "post_date": "06/28/2021 03:35:01",
          "content": "<p>Wow thats amazing for a LB like this, mabey this is the best approach if so..</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1368104,
      "author_name": "t88take",
      "author_url": "",
      "post_date": "06/28/2021 10:29:10",
      "content": "<p>gnss log has data about the movement of the vehicle. This is making my model better.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1368146,
          "author_name": "avivlevi815",
          "author_url": "",
          "post_date": "06/28/2021 11:03:22",
          "content": "<p>awsome thank you!</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1368375,
      "author_name": "avtobusbratiev",
      "author_url": "",
      "post_date": "06/28/2021 14:32:56",
      "content": "<p>Nothing but a baseline, because I still think that I havent used the baseline to the full extent yet. <br>\nI am planning to incorporate the gnss logs by the end of the week though.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1368468,
          "author_name": "avivlevi815",
          "author_url": "",
          "post_date": "06/28/2021 15:43:18",
          "content": "<p>That's very impressive, thank you for sharing!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1366691": "So far I used ground truth datasets with the baseline train/test only, I could not (yet) improve my score with any other file like the derived or raw gnss data.\n\nWhat datasets did you find beneficial?",
    "1367557": "Just used everything and threw it into a scaler and an automl tool. Idc much about this comp.",
    "1367631": "For LB score, I only using baseline.  \nAppropriate treatment of the outliers that already discussed on discussions or notebook can improve accuracy by only baseline data.",
    "1367704": "Wow thats amazing for a LB like this, mabey this is the best approach if so..",
    "1367705": "Thanks!\nIs auto ML legit in this competition?",
    "1368104": "gnss log has data about the movement of the vehicle. This is making my model better.",
    "1368146": "awsome thank you!",
    "1368375": "Nothing but a baseline, because I still think that I havent used the baseline to the full extent yet. \nI am planning to incorporate the gnss logs by the end of the week though.",
    "1368468": "That's very impressive, thank you for sharing!",
    "1369886": "Not really"
  },
  "source": "meta"
}