{
  "id": 17959,
  "title": "Evaluation formula modification",
  "url": "/competitions/second-annual-data-science-bowl/discussion/17959",
  "author_name": "",
  "post_date": "2015-12-16T15:27:38.427Z",
  "votes": 1,
  "comment_count": 1,
  "views": 883,
  "content": "<p>In case it is still possible to change evaluation formula, I would like to post the following  observation. The top and bottom slices of ventricle are hard to segment, so it is possible to obtain very accurate ejection fraction value based on measuring the contraction of well visible walls of ventricle's midsection, but have a significant error in measuring end-systolic and end-diastolic volumes (biased in the same direction). Since &quot;both the volumes and the ejection fraction are predictive of heart disease&quot;, why not add ejection fraction component to the scoring formula, which would reward precise EF measurement (e.g. by increasing N to three times number of cases and adding rows with predicted distribution of EF).</p>",
  "messages": [
    {
      "id": "101702",
      "postDate": "12/16/2015 15:27:38",
      "content": "<p>In case it is still possible to change evaluation formula, I would like to post the following  observation. The top and bottom slices of ventricle are hard to segment, so it is possible to obtain very accurate ejection fraction value based on measuring the contraction of well visible walls of ventricle's midsection, but have a significant error in measuring end-systolic and end-diastolic volumes (biased in the same direction). Since &quot;both the volumes and the ejection fraction are predictive of heart disease&quot;, why not add ejection fraction component to the scoring formula, which would reward precise EF measurement (e.g. by increasing N to three times number of cases and adding rows with predicted distribution of EF).</p>",
      "rawMarkdown": "In case it is still possible to change evaluation formula, I would like to post the following  observation. The top and bottom slices of ventricle are hard to segment, so it is possible to obtain very accurate ejection fraction value based on measuring the contraction of well visible walls of ventricle's midsection, but have a significant error in measuring end-systolic and end-diastolic volumes (biased in the same direction). Since \"both the volumes and the ejection fraction are predictive of heart disease\", why not add ejection fraction component to the scoring formula, which would reward precise EF measurement (e.g. by increasing N to three times number of cases and adding rows with predicted distribution of EF).",
      "votes": null
    },
    {
      "id": "124787",
      "postDate": "06/22/2016 11:27:21",
      "content": "<p>I think too, the accurate EF is the final aim.</p>",
      "rawMarkdown": "I think too, the accurate EF is the final aim.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 124787,
      "author_name": "luogongning",
      "author_url": "",
      "post_date": "06/22/2016 11:27:21",
      "content": "<p>I think too, the accurate EF is the final aim.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "101702": "In case it is still possible to change evaluation formula, I would like to post the following  observation. The top and bottom slices of ventricle are hard to segment, so it is possible to obtain very accurate ejection fraction value based on measuring the contraction of well visible walls of ventricle's midsection, but have a significant error in measuring end-systolic and end-diastolic volumes (biased in the same direction). Since \"both the volumes and the ejection fraction are predictive of heart disease\", why not add ejection fraction component to the scoring formula, which would reward precise EF measurement (e.g. by increasing N to three times number of cases and adding rows with predicted distribution of EF).",
    "124787": "I think too, the accurate EF is the final aim."
  },
  "source": "meta"
}