{
  "id": 178331,
  "title": "Multiple Quantile Regression With LSTM",
  "url": "/competitions/osic-pulmonary-fibrosis-progression/discussion/178331",
  "author_name": "",
  "post_date": "2020-08-29T15:24:48.106181200Z",
  "votes": 2,
  "comment_count": 4,
  "views": 0,
  "content": "<p>Hi Guys I have tried out using LSTM in my notebook but seems LSTM doesn't score well in public lb .Has anyone tried using it and got a boost???</p>\n<p>Baseline LSTM : <a href=\"https://www.kaggle.com/vineeth1999/multiple-quantile-regression-with-2-lstm\" target=\"_blank\">https://www.kaggle.com/vineeth1999/multiple-quantile-regression-with-2-lstm</a></p>",
  "messages": [
    {
      "id": "990401",
      "postDate": "08/29/2020 15:24:48",
      "content": "<p>Hi Guys I have tried out using LSTM in my notebook but seems LSTM doesn't score well in public lb .Has anyone tried using it and got a boost???</p>\n<p>Baseline LSTM : <a href=\"https://www.kaggle.com/vineeth1999/multiple-quantile-regression-with-2-lstm\" target=\"_blank\">https://www.kaggle.com/vineeth1999/multiple-quantile-regression-with-2-lstm</a></p>",
      "rawMarkdown": "Hi Guys I have tried out using LSTM in my notebook but seems LSTM doesn't score well in public lb .Has anyone tried using it and got a boost???\n\n\nBaseline LSTM : https://www.kaggle.com/vineeth1999/multiple-quantile-regression-with-2-lstm",
      "votes": null
    },
    {
      "id": "990494",
      "postDate": "08/29/2020 16:44:54",
      "content": "<p>Hi! <a href=\"https://www.kaggle.com/vineeth1999\" target=\"_blank\">@vineeth1999</a>, I haven't used LSTM, but not sure if there are enough points even to extract some information from the sequence of data.</p>",
      "rawMarkdown": "Hi! @vineeth1999, I haven't used LSTM, but not sure if there are enough points even to extract some information from the sequence of data.",
      "votes": null
    },
    {
      "id": "990800",
      "postDate": "08/29/2020 21:23:02",
      "content": "<p>You could use conditional rnn, like this <a href=\"https://github.com/philipperemy/cond_rnn\" target=\"_blank\">https://github.com/philipperemy/cond_rnn</a></p>",
      "rawMarkdown": "You could use conditional rnn, like this https://github.com/philipperemy/cond_rnn",
      "votes": null
    },
    {
      "id": "991012",
      "postDate": "08/30/2020 04:03:40",
      "content": "<p>Thanks for your insight</p>",
      "rawMarkdown": "Thanks for your insight",
      "votes": null
    },
    {
      "id": "993524",
      "postDate": "09/01/2020 02:04:23",
      "content": "<p>the time step is irregular step. some missed or no values. how about the RNN?  find a paper :<a href=\"https://www.jiqizhixin.com/articles/2020-05-20-7\" target=\"_blank\">https://www.jiqizhixin.com/articles/2020-05-20-7</a>, so difficult for me to understand ,maybe have some value for you </p>",
      "rawMarkdown": "the time step is irregular step. some missed or no values. how about the RNN?  find a paper :https://www.jiqizhixin.com/articles/2020-05-20-7, so difficult for me to understand ,maybe have some value for you",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 990494,
      "author_name": "cv13j0",
      "author_url": "",
      "post_date": "08/29/2020 16:44:54",
      "content": "<p>Hi! <a href=\"https://www.kaggle.com/vineeth1999\" target=\"_blank\">@vineeth1999</a>, I haven't used LSTM, but not sure if there are enough points even to extract some information from the sequence of data.</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 990800,
      "author_name": "eladwar",
      "author_url": "",
      "post_date": "08/29/2020 21:23:02",
      "content": "<p>You could use conditional rnn, like this <a href=\"https://github.com/philipperemy/cond_rnn\" target=\"_blank\">https://github.com/philipperemy/cond_rnn</a></p>",
      "votes": null,
      "replies": [
        {
          "id": 991012,
          "author_name": "vineeth1999",
          "author_url": "",
          "post_date": "08/30/2020 04:03:40",
          "content": "<p>Thanks for your insight</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 993524,
      "author_name": "wangyoucao1234",
      "author_url": "",
      "post_date": "09/01/2020 02:04:23",
      "content": "<p>the time step is irregular step. some missed or no values. how about the RNN?  find a paper :<a href=\"https://www.jiqizhixin.com/articles/2020-05-20-7\" target=\"_blank\">https://www.jiqizhixin.com/articles/2020-05-20-7</a>, so difficult for me to understand ,maybe have some value for you </p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "990401": "Hi Guys I have tried out using LSTM in my notebook but seems LSTM doesn't score well in public lb .Has anyone tried using it and got a boost???\n\n\nBaseline LSTM : https://www.kaggle.com/vineeth1999/multiple-quantile-regression-with-2-lstm",
    "990494": "Hi! @vineeth1999, I haven't used LSTM, but not sure if there are enough points even to extract some information from the sequence of data.",
    "990800": "You could use conditional rnn, like this https://github.com/philipperemy/cond_rnn",
    "991012": "Thanks for your insight",
    "993524": "the time step is irregular step. some missed or no values. how about the RNN?  find a paper :https://www.jiqizhixin.com/articles/2020-05-20-7, so difficult for me to understand ,maybe have some value for you"
  },
  "source": "meta"
}