{
  "id": 177200,
  "title": "Working with Tabular Data",
  "url": "/competitions/osic-pulmonary-fibrosis-progression/discussion/177200",
  "author_name": "",
  "post_date": "2020-08-25T06:18:42.609509700Z",
  "votes": 1,
  "comment_count": 3,
  "views": 0,
  "content": "<p>I'll be trying Feature Engineering on tabular data along with metadata from CT Scans. Till now I have created some lag features from the dataset as <a href=\"https://www.kaggle.com/rftexas\" target=\"_blank\">@rftexas</a>  has explained in his notebook. I'll be posting the features which give better performance and the model. For now, I am using Gradient Boosted Algorithms. Comment down below if you get some better features.</p>",
  "messages": [
    {
      "id": "984536",
      "postDate": "08/25/2020 06:18:42",
      "content": "<p>I'll be trying Feature Engineering on tabular data along with metadata from CT Scans. Till now I have created some lag features from the dataset as <a href=\"https://www.kaggle.com/rftexas\" target=\"_blank\">@rftexas</a>  has explained in his notebook. I'll be posting the features which give better performance and the model. For now, I am using Gradient Boosted Algorithms. Comment down below if you get some better features.</p>",
      "rawMarkdown": "I'll be trying Feature Engineering on tabular data along with metadata from CT Scans. Till now I have created some lag features from the dataset as @rftexas  has explained in his notebook. I'll be posting the features which give better performance and the model. For now, I am using Gradient Boosted Algorithms. Comment down below if you get some better features.",
      "votes": null
    },
    {
      "id": "985321",
      "postDate": "08/25/2020 16:36:57",
      "content": "<p>After doing some EDA found out that the patients in validation set are already in the train set. </p>",
      "rawMarkdown": "After doing some EDA found out that the patients in validation set are already in the train set.",
      "votes": null
    },
    {
      "id": "985401",
      "postDate": "08/25/2020 17:49:20",
      "content": "<p>I also have trained a simple elastic net  having week feature combined with patient getting LB score of -0.69076</p>",
      "rawMarkdown": "I also have trained a simple elastic net  having week feature combined with patient getting LB score of -0.69076",
      "votes": null
    },
    {
      "id": "985501",
      "postDate": "08/25/2020 19:00:36",
      "content": "<p>The patients in the test set are only in your notebook in the train set. If you submit your notebook a different testset will be used named 'test.csv' and corresponding images.</p>",
      "rawMarkdown": "The patients in the test set are only in your notebook in the train set. If you submit your notebook a different testset will be used named 'test.csv' and corresponding images.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 985321,
      "author_name": "pranavuikey",
      "author_url": "",
      "post_date": "08/25/2020 16:36:57",
      "content": "<p>After doing some EDA found out that the patients in validation set are already in the train set. </p>",
      "votes": null,
      "replies": [
        {
          "id": 985501,
          "author_name": "lukereijnen",
          "author_url": "",
          "post_date": "08/25/2020 19:00:36",
          "content": "<p>The patients in the test set are only in your notebook in the train set. If you submit your notebook a different testset will be used named 'test.csv' and corresponding images.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 985401,
      "author_name": "pranavuikey",
      "author_url": "",
      "post_date": "08/25/2020 17:49:20",
      "content": "<p>I also have trained a simple elastic net  having week feature combined with patient getting LB score of -0.69076</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "984536": "I'll be trying Feature Engineering on tabular data along with metadata from CT Scans. Till now I have created some lag features from the dataset as @rftexas  has explained in his notebook. I'll be posting the features which give better performance and the model. For now, I am using Gradient Boosted Algorithms. Comment down below if you get some better features.",
    "985321": "After doing some EDA found out that the patients in validation set are already in the train set.",
    "985401": "I also have trained a simple elastic net  having week feature combined with patient getting LB score of -0.69076",
    "985501": "The patients in the test set are only in your notebook in the train set. If you submit your notebook a different testset will be used named 'test.csv' and corresponding images."
  },
  "source": "meta"
}