{
  "id": 442668,
  "title": "Value of \"additional folders\"",
  "url": "/competitions/stanford-ribonanza-rna-folding/discussion/442668",
  "author_name": "",
  "post_date": "2023-09-23T17:40:11.823515800Z",
  "votes": 4,
  "comment_count": 2,
  "views": 0,
  "content": "<p>So I was doing a first pass look at the various notebooks and it appears most people are sticking with the train and test dataset csv's. I was wondering if anyone found any success it actually incorporating the additional files into a model, other then just EDA.</p>\n<p>For example the Ribonanza_bpp_files are purely from a set of simulations, but having the secondary structures has been pointed out to be of value. Open to any thoughts, comments, suggestion. </p>",
  "messages": [
    {
      "id": "2452985",
      "postDate": "09/23/2023 17:40:11",
      "content": "<p>So I was doing a first pass look at the various notebooks and it appears most people are sticking with the train and test dataset csv's. I was wondering if anyone found any success it actually incorporating the additional files into a model, other then just EDA.</p>\n<p>For example the Ribonanza_bpp_files are purely from a set of simulations, but having the secondary structures has been pointed out to be of value. Open to any thoughts, comments, suggestion. </p>",
      "rawMarkdown": "So I was doing a first pass look at the various notebooks and it appears most people are sticking with the train and test dataset csv's. I was wondering if anyone found any success it actually incorporating the additional files into a model, other then just EDA.\n\nFor example the Ribonanza_bpp_files are purely from a set of simulations, but having the secondary structures has been pointed out to be of value. Open to any thoughts, comments, suggestion.",
      "votes": null
    },
    {
      "id": "2476845",
      "postDate": "10/11/2023 01:10:43",
      "content": "<p>I am curious here too. </p>",
      "rawMarkdown": "I am curious here too.",
      "votes": null
    },
    {
      "id": "2476902",
      "postDate": "10/11/2023 02:12:19",
      "content": "<p>If these additional data are not provided in private leaderboards then it would probably be less meaningful to analyze?</p>",
      "rawMarkdown": "If these additional data are not provided in private leaderboards then it would probably be less meaningful to analyze?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2476845,
      "author_name": "ratnesh1729",
      "author_url": "",
      "post_date": "10/11/2023 01:10:43",
      "content": "<p>I am curious here too. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 2476902,
      "author_name": "ratnesh1729",
      "author_url": "",
      "post_date": "10/11/2023 02:12:19",
      "content": "<p>If these additional data are not provided in private leaderboards then it would probably be less meaningful to analyze?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2452985": "So I was doing a first pass look at the various notebooks and it appears most people are sticking with the train and test dataset csv's. I was wondering if anyone found any success it actually incorporating the additional files into a model, other then just EDA.\n\nFor example the Ribonanza_bpp_files are purely from a set of simulations, but having the secondary structures has been pointed out to be of value. Open to any thoughts, comments, suggestion.",
    "2476845": "I am curious here too.",
    "2476902": "If these additional data are not provided in private leaderboards then it would probably be less meaningful to analyze?"
  },
  "source": "meta"
}