{
  "id": 2414,
  "title": "Algorithm or Result or Both?",
  "url": "/competitions/predict-closed-questions-on-stack-overflow/discussion/2414",
  "author_name": "",
  "post_date": "2012-08-22T20:03:47.157Z",
  "votes": null,
  "comment_count": 1,
  "views": 2333,
  "content": "<p>I'm brand new to kaggle, inspired by the stackexchange blog post and the prospect of making stackoverflow better.</p>\r\n<p>I get the overall goal... but I'm a bit confused by the contest and what the &quot;gist&quot; of this challenge is.</p>\r\n<p>&nbsp;</p>\r\n<p>Is the idea to come up with a program that can be run against the question content before it is accepted and posted to the front page? (e.g. the Algorithm) or just generate the &quot;ranking&quot; of a particular known question? (e.g. the Result)... or both?</p>\r\n<p>I'd like to participate (even if just to see if I can accomplish a respectable result) but I'd also like to enter the contest if I can.</p>",
  "messages": [
    {
      "id": "13338",
      "postDate": "08/22/2012 20:03:47",
      "content": "<p>I'm brand new to kaggle, inspired by the stackexchange blog post and the prospect of making stackoverflow better.</p>\r\n<p>I get the overall goal... but I'm a bit confused by the contest and what the &quot;gist&quot; of this challenge is.</p>\r\n<p>&nbsp;</p>\r\n<p>Is the idea to come up with a program that can be run against the question content before it is accepted and posted to the front page? (e.g. the Algorithm) or just generate the &quot;ranking&quot; of a particular known question? (e.g. the Result)... or both?</p>\r\n<p>I'd like to participate (even if just to see if I can accomplish a respectable result) but I'd also like to enter the contest if I can.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "13342",
      "postDate": "08/22/2012 22:53:11",
      "content": "<p>The idea is to take in a question as it exists prior to posting (so you know it's text, and some metadata about it's owner) and predict the likelihood it will remain open, and if not how likely it was closed as each close reason.</p>\r\n<p>So we're looking for the Algorithm.</p>\r\n<p>The way Kaggle works is to take the Results your Algorithm produces for some test data for the leaderboard. Prior to selecting winners we'll require everyone submit their Algorithms to be run against fresh data.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 13342,
      "author_name": "kevinmontrose",
      "author_url": "",
      "post_date": "08/22/2012 22:53:11",
      "content": "<p>The idea is to take in a question as it exists prior to posting (so you know it's text, and some metadata about it's owner) and predict the likelihood it will remain open, and if not how likely it was closed as each close reason.</p>\r\n<p>So we're looking for the Algorithm.</p>\r\n<p>The way Kaggle works is to take the Results your Algorithm produces for some test data for the leaderboard. Prior to selecting winners we'll require everyone submit their Algorithms to be run against fresh data.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "13338": "",
    "13342": ""
  },
  "source": "meta"
}