{
  "id": 10913,
  "title": "Genetic Programming",
  "url": "/competitions/seizure-prediction/discussion/10913",
  "author_name": "",
  "post_date": "2014-11-12T14:51:59.050Z",
  "votes": null,
  "comment_count": 1,
  "views": 706,
  "content": "<p>Is anyone&nbsp;else attempting to classify the signals by means of genetic programming?</p>",
  "messages": [
    {
      "id": "57912",
      "postDate": "11/12/2014 14:51:59",
      "content": "<p>Is anyone&nbsp;else attempting to classify the signals by means of genetic programming?</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "58247",
      "postDate": "11/18/2014 00:45:03",
      "content": "<p>I used genetic algorithm for feature selection to find better subsets of the data to train with. This appeared to strongly improve performance for Dog_3 and Dog_4, little to no effect on Dog_1 and Dog_2, negative effect for Patient_1 and Patient_2. I didn't check Dog_5. This was based on leaderboard scores.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 58247,
      "author_name": "michaelhills",
      "author_url": "",
      "post_date": "11/18/2014 00:45:03",
      "content": "<p>I used genetic algorithm for feature selection to find better subsets of the data to train with. This appeared to strongly improve performance for Dog_3 and Dog_4, little to no effect on Dog_1 and Dog_2, negative effect for Patient_1 and Patient_2. I didn't check Dog_5. This was based on leaderboard scores.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "57912": "",
    "58247": ""
  },
  "source": "meta"
}