{
  "id": 182730,
  "title": "How do you concatinate tabular data and image data",
  "url": "/competitions/osic-pulmonary-fibrosis-progression/discussion/182730",
  "author_name": "",
  "post_date": "2020-09-14T05:50:09.127529200Z",
  "votes": 4,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I think it is the key to have good score to concatinate tabular and image data. However this is the most difficult part in this competition. If you don't mind, please share your idea to solve this part.</p>",
  "messages": [
    {
      "id": "1009591",
      "postDate": "09/14/2020 05:50:09",
      "content": "<p>I think it is the key to have good score to concatinate tabular and image data. However this is the most difficult part in this competition. If you don't mind, please share your idea to solve this part.</p>",
      "rawMarkdown": "I think it is the key to have good score to concatinate tabular and image data. However this is the most difficult part in this competition. If you don't mind, please share your idea to solve this part.",
      "votes": null
    },
    {
      "id": "1010252",
      "postDate": "09/14/2020 16:06:59",
      "content": "<p>Good question. I’m on mobile so it’s hard to link to specific notebooks, but one possible idea is to use a convolutional network to extract features from the images, and a dense layer to extract features from the tabular data, and concatenate both before proceeding with training dense layers on the final, higher-level layers. </p>",
      "rawMarkdown": "Good question. I’m on mobile so it’s hard to link to specific notebooks, but one possible idea is to use a convolutional network to extract features from the images, and a dense layer to extract features from the tabular data, and concatenate both before proceeding with training dense layers on the final, higher-level layers.",
      "votes": null
    },
    {
      "id": "1011288",
      "postDate": "09/15/2020 11:15:55",
      "content": "<p>Thank you for your advise!!<br>\nI will try to make a model based on your idea!</p>",
      "rawMarkdown": "Thank you for your advise!!\nI will try to make a model based on your idea!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1010252,
      "author_name": "douglaskgaraujo",
      "author_url": "",
      "post_date": "09/14/2020 16:06:59",
      "content": "<p>Good question. I’m on mobile so it’s hard to link to specific notebooks, but one possible idea is to use a convolutional network to extract features from the images, and a dense layer to extract features from the tabular data, and concatenate both before proceeding with training dense layers on the final, higher-level layers. </p>",
      "votes": null,
      "replies": [
        {
          "id": 1011288,
          "author_name": "taikiishii",
          "author_url": "",
          "post_date": "09/15/2020 11:15:55",
          "content": "<p>Thank you for your advise!!<br>\nI will try to make a model based on your idea!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "1009591": "I think it is the key to have good score to concatinate tabular and image data. However this is the most difficult part in this competition. If you don't mind, please share your idea to solve this part.",
    "1010252": "Good question. I’m on mobile so it’s hard to link to specific notebooks, but one possible idea is to use a convolutional network to extract features from the images, and a dense layer to extract features from the tabular data, and concatenate both before proceeding with training dense layers on the final, higher-level layers.",
    "1011288": "Thank you for your advise!!\nI will try to make a model based on your idea!"
  },
  "source": "meta"
}