{
  "id": 5638,
  "title": "Welcome",
  "url": "/competitions/expedia-personalized-sort/discussion/5638",
  "author_name": "",
  "post_date": "2013-09-03T17:08:05.393Z",
  "votes": null,
  "comment_count": 3,
  "views": 2373,
  "content": "<p>We're excited to launch this recommendation engine competition with <a href=\"http://www.expedia.com/\">Expedia</a>&nbsp;and <a href=\"http://icdm2013.rutgers.edu/\">ICDM 2013</a>!</p>\n<p>To get started, download the data and sample submissions from the<a href=\"https://www.kaggle.com/c/expedia-personalized-sort/data\"> data page</a>. Starter Python code to create the benchmarks is <a href=\"https://github.com/benhamner/ExpediaPersonalizedSortCompetition\">available on Github</a>.</p>\n<p>Good luck on the competition, and we hope you have fun with it!</p>",
  "messages": [
    {
      "id": "30070",
      "postDate": "09/03/2013 17:08:05",
      "content": "<p>We're excited to launch this recommendation engine competition with <a href=\"http://www.expedia.com/\">Expedia</a>&nbsp;and <a href=\"http://icdm2013.rutgers.edu/\">ICDM 2013</a>!</p>\n<p>To get started, download the data and sample submissions from the<a href=\"https://www.kaggle.com/c/expedia-personalized-sort/data\"> data page</a>. Starter Python code to create the benchmarks is <a href=\"https://github.com/benhamner/ExpediaPersonalizedSortCompetition\">available on Github</a>.</p>\n<p>Good luck on the competition, and we hope you have fun with it!</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "30082",
      "postDate": "09/03/2013 20:06:03",
      "content": "<p>the readme says it will take approx 5mins to run but on my Mac,&nbsp;with 8GB of RAM and 4 cores at 2.7GHz, it takes forever to &quot;predict&quot;, however, the training is quite fast :D</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "30187",
      "postDate": "09/04/2013 20:37:58",
      "content": "<p>OK, something I think that should be release along with dataset for every contest: source code for the evaluation function please.</p>\n<p><span style=\"line-height: 1.4\">And/or some dummy data set with scores calculated by the actual leaderboard scoring function just for our peace of mind.</span></p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "30405",
      "postDate": "09/07/2013 04:03:38",
      "content": "<p>The source code for the evaluation metric is below. (I'm providing it for reference - we don't expect you to be able to run it directly at this point since it has references that aren't included).</p>\n<p>Edit: copied it onto the wiki, since the formatting didn't work for that length of a text block:&nbsp;https://www.kaggle.com/wiki/NormalizedDiscountedCumulativeGain</p>\n<p>I also attached a zip file with a test case to the wiki.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 30082,
      "author_name": "abhishek",
      "author_url": "",
      "post_date": "09/03/2013 20:06:03",
      "content": "<p>the readme says it will take approx 5mins to run but on my Mac,&nbsp;with 8GB of RAM and 4 cores at 2.7GHz, it takes forever to &quot;predict&quot;, however, the training is quite fast :D</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 30187,
      "author_name": "byang1",
      "author_url": "",
      "post_date": "09/04/2013 20:37:58",
      "content": "<p>OK, something I think that should be release along with dataset for every contest: source code for the evaluation function please.</p>\n<p><span style=\"line-height: 1.4\">And/or some dummy data set with scores calculated by the actual leaderboard scoring function just for our peace of mind.</span></p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 30405,
      "author_name": "benhamner",
      "author_url": "",
      "post_date": "09/07/2013 04:03:38",
      "content": "<p>The source code for the evaluation metric is below. (I'm providing it for reference - we don't expect you to be able to run it directly at this point since it has references that aren't included).</p>\n<p>Edit: copied it onto the wiki, since the formatting didn't work for that length of a text block:&nbsp;https://www.kaggle.com/wiki/NormalizedDiscountedCumulativeGain</p>\n<p>I also attached a zip file with a test case to the wiki.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "30070": "",
    "30082": "",
    "30187": "",
    "30405": ""
  },
  "source": "meta"
}