{
  "id": 416036,
  "title": "Neural Network Domination",
  "url": "/competitions/tlvmc-parkinsons-freezing-gait-prediction/discussion/416036",
  "author_name": "",
  "post_date": "2023-06-09T08:36:54.417945200Z",
  "votes": 2,
  "comment_count": 3,
  "views": 0,
  "content": "<p>After a huge shakeup, it seems that all the top models (and the models that managed to survive the shakeup) were based on neural networks. Just a general observation. I might be wrong thou. <br>\nFor instance, public notebook by <a href=\"https://www.kaggle.com/coderrkj\" target=\"_blank\">@coderrkj</a> gave higher private score of 0.298 while notebooks using LGBM regressor suffered a huge shakeup. </p>",
  "messages": [
    {
      "id": "2293525",
      "postDate": "06/09/2023 08:36:54",
      "content": "<p>After a huge shakeup, it seems that all the top models (and the models that managed to survive the shakeup) were based on neural networks. Just a general observation. I might be wrong thou. <br>\nFor instance, public notebook by <a href=\"https://www.kaggle.com/coderrkj\" target=\"_blank\">@coderrkj</a> gave higher private score of 0.298 while notebooks using LGBM regressor suffered a huge shakeup. </p>",
      "rawMarkdown": "After a huge shakeup, it seems that all the top models (and the models that managed to survive the shakeup) were based on neural networks. Just a general observation. I might be wrong thou. \nFor instance, public notebook by @coderrkj gave higher private score of 0.298 while notebooks using LGBM regressor suffered a huge shakeup.",
      "votes": null
    },
    {
      "id": "2293577",
      "postDate": "06/09/2023 09:41:04",
      "content": "<p>My best lgb submission can achieve private :0.337+public :0.45(around 15st place).<br>\nBut  I choose the higher public one finally(private:0.319+public:0.47) </p>",
      "rawMarkdown": "My best lgb submission can achieve private :0.337+public :0.45(around 15st place).\nBut  I choose the higher public one finally(private:0.319+public:0.47)",
      "votes": null
    },
    {
      "id": "2293679",
      "postDate": "06/09/2023 11:58:09",
      "content": "<p>Oh! I see. Btw, congratulations on the silver medal! :)<br>\nQuick question though, did you use parquet files as well, or just the files in the train folder? </p>",
      "rawMarkdown": "Oh! I see. Btw, congratulations on the silver medal! :)\nQuick question though, did you use parquet files as well, or just the files in the train folder?",
      "votes": null
    },
    {
      "id": "2293994",
      "postDate": "06/09/2023 16:37:01",
      "content": "<p>I didn't use the unlabeled data and notype data</p>",
      "rawMarkdown": "I didn't use the unlabeled data and notype data",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2293577,
      "author_name": "xzj19013742",
      "author_url": "",
      "post_date": "06/09/2023 09:41:04",
      "content": "<p>My best lgb submission can achieve private :0.337+public :0.45(around 15st place).<br>\nBut  I choose the higher public one finally(private:0.319+public:0.47) </p>",
      "votes": null,
      "replies": [
        {
          "id": 2293679,
          "author_name": "ahsuna123",
          "author_url": "",
          "post_date": "06/09/2023 11:58:09",
          "content": "<p>Oh! I see. Btw, congratulations on the silver medal! :)<br>\nQuick question though, did you use parquet files as well, or just the files in the train folder? </p>",
          "votes": null,
          "replies": [
            {
              "id": 2293994,
              "author_name": "xzj19013742",
              "author_url": "",
              "post_date": "06/09/2023 16:37:01",
              "content": "<p>I didn't use the unlabeled data and notype data</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2293525": "After a huge shakeup, it seems that all the top models (and the models that managed to survive the shakeup) were based on neural networks. Just a general observation. I might be wrong thou. \nFor instance, public notebook by @coderrkj gave higher private score of 0.298 while notebooks using LGBM regressor suffered a huge shakeup.",
    "2293577": "My best lgb submission can achieve private :0.337+public :0.45(around 15st place).\nBut  I choose the higher public one finally(private:0.319+public:0.47)",
    "2293679": "Oh! I see. Btw, congratulations on the silver medal! :)\nQuick question though, did you use parquet files as well, or just the files in the train folder?",
    "2293994": "I didn't use the unlabeled data and notype data"
  },
  "source": "meta"
}