{
  "id": 420964,
  "title": "Gold Medalist Solutions",
  "url": "/competitions/predict-student-performance-from-game-play/discussion/420964",
  "author_name": "",
  "post_date": "2023-07-03T11:43:33.683485900Z",
  "votes": 7,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I would like to thank the organizers and Kaggle for their efforts in organizing this competition.<br>\nI would also like to thank the many kagglers who published their solutions immediately after the competition.<br>\nAs this is my first time participating in the competition, I am trying to learn a lot from them, whom I respect very much.<br>\nHere is a brief list of what I learned from them. (except for NN, as I do not understand NN.)</p>\n<ul>\n<li>trust CV: I was only concerned about Public LB.</li>\n<li>use of raw data(<a href=\"url\" target=\"_blank\">https://fielddaylab.wisc.edu/opengamedata/</a>): score improved by about 0.002</li>\n<li>model: Top 1, 3, and 4 are all GBDT + NN ensemble models</li>\n<li>effective feature engineering: e.g., event_name + name, answer time, etc.</li>\n</ul>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F9361058%2Fee9214cc87dc2293e2013aff8043930e%2FPSP.PNG?generation=1688384376349866&amp;alt=media\" alt=\"\"></p>\n<p>I could learn many other valid approaches from them.<br>\nI would like to express my gratitude again and will use them in my future studies.<br>\nIf you could give me further advice, I would be happy to do so.<br>\nThank you for reading.</p>",
  "messages": [
    {
      "id": "2328154",
      "postDate": "07/03/2023 11:43:33",
      "content": "<p>I would like to thank the organizers and Kaggle for their efforts in organizing this competition.<br>\nI would also like to thank the many kagglers who published their solutions immediately after the competition.<br>\nAs this is my first time participating in the competition, I am trying to learn a lot from them, whom I respect very much.<br>\nHere is a brief list of what I learned from them. (except for NN, as I do not understand NN.)</p>\n<ul>\n<li>trust CV: I was only concerned about Public LB.</li>\n<li>use of raw data(<a href=\"url\" target=\"_blank\">https://fielddaylab.wisc.edu/opengamedata/</a>): score improved by about 0.002</li>\n<li>model: Top 1, 3, and 4 are all GBDT + NN ensemble models</li>\n<li>effective feature engineering: e.g., event_name + name, answer time, etc.</li>\n</ul>\n<p><img src=\"https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F9361058%2Fee9214cc87dc2293e2013aff8043930e%2FPSP.PNG?generation=1688384376349866&amp;alt=media\" alt=\"\"></p>\n<p>I could learn many other valid approaches from them.<br>\nI would like to express my gratitude again and will use them in my future studies.<br>\nIf you could give me further advice, I would be happy to do so.<br>\nThank you for reading.</p>",
      "rawMarkdown": "I would like to thank the organizers and Kaggle for their efforts in organizing this competition.\nI would also like to thank the many kagglers who published their solutions immediately after the competition.\nAs this is my first time participating in the competition, I am trying to learn a lot from them, whom I respect very much.\nHere is a brief list of what I learned from them. (except for NN, as I do not understand NN.)\n\n- trust CV: I was only concerned about Public LB.\n- use of raw data([https://fielddaylab.wisc.edu/opengamedata/](url)): score improved by about 0.002\n- model: Top 1, 3, and 4 are all GBDT + NN ensemble models\n- effective feature engineering: e.g., event_name + name, answer time, etc.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F9361058%2Fee9214cc87dc2293e2013aff8043930e%2FPSP.PNG?generation=1688384376349866&alt=media)\n\nI could learn many other valid approaches from them.\nI would like to express my gratitude again and will use them in my future studies.\nIf you could give me further advice, I would be happy to do so.\nThank you for reading.",
      "votes": null
    },
    {
      "id": "2329311",
      "postDate": "07/04/2023 07:41:33",
      "content": "<p>Thank you for the summary!🎉  I learned a lot from the competition too!</p>",
      "rawMarkdown": "Thank you for the summary!🎉  I learned a lot from the competition too!",
      "votes": null
    },
    {
      "id": "2329657",
      "postDate": "07/04/2023 12:13:47",
      "content": "<p>Thanks for your positive comments.<br>\nAnd congratulations on your silver medal!</p>",
      "rawMarkdown": "Thanks for your positive comments.\nAnd congratulations on your silver medal!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2329311,
      "author_name": "kaggleaau",
      "author_url": "",
      "post_date": "07/04/2023 07:41:33",
      "content": "<p>Thank you for the summary!🎉  I learned a lot from the competition too!</p>",
      "votes": null,
      "replies": [
        {
          "id": 2329657,
          "author_name": "inoway",
          "author_url": "",
          "post_date": "07/04/2023 12:13:47",
          "content": "<p>Thanks for your positive comments.<br>\nAnd congratulations on your silver medal!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2328154": "I would like to thank the organizers and Kaggle for their efforts in organizing this competition.\nI would also like to thank the many kagglers who published their solutions immediately after the competition.\nAs this is my first time participating in the competition, I am trying to learn a lot from them, whom I respect very much.\nHere is a brief list of what I learned from them. (except for NN, as I do not understand NN.)\n\n- trust CV: I was only concerned about Public LB.\n- use of raw data([https://fielddaylab.wisc.edu/opengamedata/](url)): score improved by about 0.002\n- model: Top 1, 3, and 4 are all GBDT + NN ensemble models\n- effective feature engineering: e.g., event_name + name, answer time, etc.\n\n![](https://www.googleapis.com/download/storage/v1/b/kaggle-forum-message-attachments/o/inbox%2F9361058%2Fee9214cc87dc2293e2013aff8043930e%2FPSP.PNG?generation=1688384376349866&alt=media)\n\nI could learn many other valid approaches from them.\nI would like to express my gratitude again and will use them in my future studies.\nIf you could give me further advice, I would be happy to do so.\nThank you for reading.",
    "2329311": "Thank you for the summary!🎉  I learned a lot from the competition too!",
    "2329657": "Thanks for your positive comments.\nAnd congratulations on your silver medal!"
  },
  "source": "meta"
}