{
  "id": 75457,
  "title": "Matthews Correlation Coefficient",
  "url": "/competitions/vsb-power-line-fault-detection/discussion/75457",
  "author_name": "Panchajanya Banerjee (Pancham)",
  "post_date": "2018-12-21T20:48:42.315000",
  "votes": 7,
  "comment_count": 0,
  "views": 0,
  "content": "<p>This article gives some valuable intuition about the evaluation metric for this competition::</p>\n\n<p><a href=\"https://lettier.github.io/posts/2016-08-05-matthews-correlation-coefficient.html\">https://lettier.github.io/posts/2016-08-05-matthews-correlation-coefficient.html</a></p>\n\n<p>Obviously, the <a href=\"https://en.wikipedia.org/wiki/Matthews_correlation_coefficient\">wikipedia article</a> which was linked in the competition description is also rather helpful. </p>\n\n<p>From what I gather, a good metric especially when classifying unbalanced data such as we have in our case (most power-lines are okay, a crucial few are not)</p>\n\n<p>Further readings::\n* <a href=\"https://www.bioinfopublication.org/files/articles/2_1_1_JMLT.pdf\">https://www.bioinfopublication.org/files/articles/2_1_1_JMLT.pdf</a>\n* <a href=\"http://standardwisdom.com/softwarejournal/2010/06/matthews-correlation-coefficient/\">http://standardwisdom.com/softwarejournal/2010/06/matthews-correlation-coefficient/</a></p>",
  "messages": [
    {
      "id": 443546,
      "postDate": "2018-12-21T20:48:42.317Z",
      "content": "<p>This article gives some valuable intuition about the evaluation metric for this competition::</p>\n\n<p><a href=\"https://lettier.github.io/posts/2016-08-05-matthews-correlation-coefficient.html\">https://lettier.github.io/posts/2016-08-05-matthews-correlation-coefficient.html</a></p>\n\n<p>Obviously, the <a href=\"https://en.wikipedia.org/wiki/Matthews_correlation_coefficient\">wikipedia article</a> which was linked in the competition description is also rather helpful. </p>\n\n<p>From what I gather, a good metric especially when classifying unbalanced data such as we have in our case (most power-lines are okay, a crucial few are not)</p>\n\n<p>Further readings::\n* <a href=\"https://www.bioinfopublication.org/files/articles/2_1_1_JMLT.pdf\">https://www.bioinfopublication.org/files/articles/2_1_1_JMLT.pdf</a>\n* <a href=\"http://standardwisdom.com/softwarejournal/2010/06/matthews-correlation-coefficient/\">http://standardwisdom.com/softwarejournal/2010/06/matthews-correlation-coefficient/</a></p>",
      "rawMarkdown": "This article gives some valuable intuition about the evaluation metric for this competition::\n\nhttps://lettier.github.io/posts/2016-08-05-matthews-correlation-coefficient.html\n\nObviously, the [wikipedia article][1] which was linked in the competition description is also rather helpful. \n\nFrom what I gather, a good metric especially when classifying unbalanced data such as we have in our case (most power-lines are okay, a crucial few are not)\n\nFurther readings::\n* https://www.bioinfopublication.org/files/articles/2_1_1_JMLT.pdf\n* http://standardwisdom.com/softwarejournal/2010/06/matthews-correlation-coefficient/\n\n\n  [1]: https://en.wikipedia.org/wiki/Matthews_correlation_coefficient",
      "votes": 7
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "443546": "This article gives some valuable intuition about the evaluation metric for this competition::\n\nhttps://lettier.github.io/posts/2016-08-05-matthews-correlation-coefficient.html\n\nObviously, the [wikipedia article][1] which was linked in the competition description is also rather helpful. \n\nFrom what I gather, a good metric especially when classifying unbalanced data such as we have in our case (most power-lines are okay, a crucial few are not)\n\nFurther readings::\n* https://www.bioinfopublication.org/files/articles/2_1_1_JMLT.pdf\n* http://standardwisdom.com/softwarejournal/2010/06/matthews-correlation-coefficient/\n\n\n  [1]: https://en.wikipedia.org/wiki/Matthews_correlation_coefficient"
  }
}