{
  "id": 90072,
  "title": "Playing with the peaks",
  "url": "/competitions/LANL-Earthquake-Prediction/discussion/90072",
  "author_name": "",
  "post_date": "2019-04-19T22:17:42.195552900Z",
  "votes": 5,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Features are also about playing with the peaks. No need to write a function, there is already a library <a href=\"https://docs.scipy.org/doc/scipy/reference/signal.html\">https://docs.scipy.org/doc/scipy/reference/signal.html</a></p>\n\n<p>For example:</p>\n\n<p><a href=\"https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.find_peaks.html\">https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.find_peaks.html</a></p>\n\n<p>peaks, properties = signal.find_peaks(y)</p>\n\n<p>widths = signal.peak_widths(y, peaks)</p>\n\n<p>etc etc....</p>\n\n<p>I think a lot of useful features (even ones that we did not think about) can be derived with this lib</p>",
  "messages": [
    {
      "id": "519976",
      "postDate": "04/19/2019 22:17:42",
      "content": "<p>Features are also about playing with the peaks. No need to write a function, there is already a library <a href=\"https://docs.scipy.org/doc/scipy/reference/signal.html\">https://docs.scipy.org/doc/scipy/reference/signal.html</a></p>\n\n<p>For example:</p>\n\n<p><a href=\"https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.find_peaks.html\">https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.find_peaks.html</a></p>\n\n<p>peaks, properties = signal.find_peaks(y)</p>\n\n<p>widths = signal.peak_widths(y, peaks)</p>\n\n<p>etc etc....</p>\n\n<p>I think a lot of useful features (even ones that we did not think about) can be derived with this lib</p>",
      "rawMarkdown": "Features are also about playing with the peaks. No need to write a function, there is already a library https://docs.scipy.org/doc/scipy/reference/signal.html\n\nFor example:\n\nhttps://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.find_peaks.html\n\npeaks, properties = signal.find_peaks(y)\n\n widths = signal.peak_widths(y, peaks)\n\netc etc....\n\nI think a lot of useful features (even ones that we did not think about) can be derived with this lib",
      "votes": null
    },
    {
      "id": "520070",
      "postDate": "04/20/2019 05:03:01",
      "content": "<p>I'm agree with you, that peaks can be very useful features. But, in my opinion, <code>scipy</code> realization are not optimal for this task. Own realisation of peaks finder can be more faster, give better result and be more flexibale. For example you can watch <a href=\"https://www.kaggle.com/mark4h/vsb-1st-place-solution\">decision of the winner of the VSB Power Line Fault Detection competion</a>.</p>",
      "rawMarkdown": "I'm agree with you, that peaks can be very useful features. But, in my opinion, ```scipy``` realization are not optimal for this task. Own realisation of peaks finder can be more faster, give better result and be more flexibale. For example you can watch [decision of the winner of the VSB Power Line Fault Detection competion](https://www.kaggle.com/mark4h/vsb-1st-place-solution).",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 520070,
      "author_name": "miklgr500",
      "author_url": "",
      "post_date": "04/20/2019 05:03:01",
      "content": "<p>I'm agree with you, that peaks can be very useful features. But, in my opinion, <code>scipy</code> realization are not optimal for this task. Own realisation of peaks finder can be more faster, give better result and be more flexibale. For example you can watch <a href=\"https://www.kaggle.com/mark4h/vsb-1st-place-solution\">decision of the winner of the VSB Power Line Fault Detection competion</a>.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "519976": "Features are also about playing with the peaks. No need to write a function, there is already a library https://docs.scipy.org/doc/scipy/reference/signal.html\n\nFor example:\n\nhttps://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.find_peaks.html\n\npeaks, properties = signal.find_peaks(y)\n\n widths = signal.peak_widths(y, peaks)\n\netc etc....\n\nI think a lot of useful features (even ones that we did not think about) can be derived with this lib",
    "520070": "I'm agree with you, that peaks can be very useful features. But, in my opinion, ```scipy``` realization are not optimal for this task. Own realisation of peaks finder can be more faster, give better result and be more flexibale. For example you can watch [decision of the winner of the VSB Power Line Fault Detection competion](https://www.kaggle.com/mark4h/vsb-1st-place-solution)."
  },
  "source": "meta"
}