{
  "id": 77654,
  "title": "Is anybody else experimented with FFT generated features?",
  "url": "/competitions/LANL-Earthquake-Prediction/discussion/77654",
  "author_name": "",
  "post_date": "2019-01-15T09:18:31.436794Z",
  "votes": 2,
  "comment_count": 2,
  "views": 0,
  "content": "<p>I started to experiment with FFT-generated features. \nI used first few harmonics (both absolute values and real components) generated from the segment.  But adding these features the results went actually worst.\nSo, is anybody else experimented with FFT-generated features?</p>",
  "messages": [
    {
      "id": "456170",
      "postDate": "01/15/2019 09:18:31",
      "content": "<p>I started to experiment with FFT-generated features. \nI used first few harmonics (both absolute values and real components) generated from the segment.  But adding these features the results went actually worst.\nSo, is anybody else experimented with FFT-generated features?</p>",
      "rawMarkdown": "I started to experiment with FFT-generated features. \nI used first few harmonics (both absolute values and real components) generated from the segment.  But adding these features the results went actually worst.\nSo, is anybody else experimented with FFT-generated features?",
      "votes": null
    },
    {
      "id": "456501",
      "postDate": "01/15/2019 23:45:23",
      "content": "<p>Yes, I have calculated some statistical features from real and imaginary part of FFT and my score has improved.</p>",
      "rawMarkdown": "Yes, I have calculated some statistical features from real and imaginary part of FFT and my score has improved.",
      "votes": null
    },
    {
      "id": "456644",
      "postDate": "01/16/2019 08:30:07",
      "content": "<p>I thought, looking to my training and validation score that this will work for me too; I think that it is something wrong with my models (not my engineered features).</p>",
      "rawMarkdown": "I thought, looking to my training and validation score that this will work for me too; I think that it is something wrong with my models (not my engineered features).",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 456501,
      "author_name": "unmonoqueteclea",
      "author_url": "",
      "post_date": "01/15/2019 23:45:23",
      "content": "<p>Yes, I have calculated some statistical features from real and imaginary part of FFT and my score has improved.</p>",
      "votes": null,
      "replies": [
        {
          "id": 456644,
          "author_name": "gpreda",
          "author_url": "",
          "post_date": "01/16/2019 08:30:07",
          "content": "<p>I thought, looking to my training and validation score that this will work for me too; I think that it is something wrong with my models (not my engineered features).</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "456170": "I started to experiment with FFT-generated features. \nI used first few harmonics (both absolute values and real components) generated from the segment.  But adding these features the results went actually worst.\nSo, is anybody else experimented with FFT-generated features?",
    "456501": "Yes, I have calculated some statistical features from real and imaginary part of FFT and my score has improved.",
    "456644": "I thought, looking to my training and validation score that this will work for me too; I think that it is something wrong with my models (not my engineered features)."
  },
  "source": "meta"
}