{
  "id": 202369,
  "title": "TabNet and model explanation ",
  "url": "/competitions/riiid-test-answer-prediction/discussion/202369",
  "author_name": "",
  "post_date": "2020-12-09T17:06:33.490522400Z",
  "votes": 3,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I wanted to show how to simply use TabNet and more importantly how you can use the attention mechanism to understand better the model and explain single level prediction.</p>\n<p>Here is the notebook: <a href=\"https://www.kaggle.com/optimo/tabnet-with-loop-feature-engineering-explained\" target=\"_blank\">https://www.kaggle.com/optimo/tabnet-with-loop-feature-engineering-explained</a></p>\n<p>Would be happy to get feedbacks on this topic.</p>\n<p>Cheers!</p>",
  "messages": [
    {
      "id": "1107404",
      "postDate": "12/09/2020 17:06:33",
      "content": "<p>I wanted to show how to simply use TabNet and more importantly how you can use the attention mechanism to understand better the model and explain single level prediction.</p>\n<p>Here is the notebook: <a href=\"https://www.kaggle.com/optimo/tabnet-with-loop-feature-engineering-explained\" target=\"_blank\">https://www.kaggle.com/optimo/tabnet-with-loop-feature-engineering-explained</a></p>\n<p>Would be happy to get feedbacks on this topic.</p>\n<p>Cheers!</p>",
      "rawMarkdown": "I wanted to show how to simply use TabNet and more importantly how you can use the attention mechanism to understand better the model and explain single level prediction.\n\nHere is the notebook: https://www.kaggle.com/optimo/tabnet-with-loop-feature-engineering-explained\n\nWould be happy to get feedbacks on this topic.\n\nCheers!",
      "votes": null
    },
    {
      "id": "1107556",
      "postDate": "12/09/2020 19:15:37",
      "content": "<p>Thanks <a href=\"https://www.kaggle.com/optimo\" target=\"_blank\">@optimo</a>. Great to see you share your tabnet knowledge here again after MOA.</p>",
      "rawMarkdown": "Thanks @optimo. Great to see you share your tabnet knowledge here again after MOA.",
      "votes": null
    },
    {
      "id": "1136509",
      "postDate": "01/03/2021 06:43:09",
      "content": "<p>Thank you for your great sharing!</p>",
      "rawMarkdown": "Thank you for your great sharing!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1107556,
      "author_name": "watzisname",
      "author_url": "",
      "post_date": "12/09/2020 19:15:37",
      "content": "<p>Thanks <a href=\"https://www.kaggle.com/optimo\" target=\"_blank\">@optimo</a>. Great to see you share your tabnet knowledge here again after MOA.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 1136509,
      "author_name": "xstargate",
      "author_url": "",
      "post_date": "01/03/2021 06:43:09",
      "content": "<p>Thank you for your great sharing!</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1107404": "I wanted to show how to simply use TabNet and more importantly how you can use the attention mechanism to understand better the model and explain single level prediction.\n\nHere is the notebook: https://www.kaggle.com/optimo/tabnet-with-loop-feature-engineering-explained\n\nWould be happy to get feedbacks on this topic.\n\nCheers!",
    "1107556": "Thanks @optimo. Great to see you share your tabnet knowledge here again after MOA.",
    "1136509": "Thank you for your great sharing!"
  },
  "source": "meta"
}