{
  "id": 2107,
  "title": "How do we build a validation set when the training set is unranked?",
  "url": "/competitions/FacebookRecruiting/discussion/2107",
  "author_name": "",
  "post_date": "2012-06-29T03:21:46.507Z",
  "votes": null,
  "comment_count": 1,
  "views": 2809,
  "content": "<p>Perhaps I've missed something, how are we supposed to build a validation set when the training isn't ranked? Moreover, what defines a ranking in this problem? I've never seen a social network where the friends/followers were ranked by the target user.</p>\r\n<p>How are we supposed to match the results of Facebook's blackbox classifier when we don't have numerical scores to train from.&nbsp;</p>",
  "messages": [
    {
      "id": "12012",
      "postDate": "06/29/2012 03:21:46",
      "content": "<p>Perhaps I've missed something, how are we supposed to build a validation set when the training isn't ranked? Moreover, what defines a ranking in this problem? I've never seen a social network where the friends/followers were ranked by the target user.</p>\r\n<p>How are we supposed to match the results of Facebook's blackbox classifier when we don't have numerical scores to train from.&nbsp;</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "12017",
      "postDate": "06/29/2012 12:12:02",
      "content": "<p>I think you are misunderstanding how the scoring for this competition works.&nbsp; Have a look at this thread: https://www.kaggle.com/c/FacebookRecruiting/forums/t/2002/alternate-explanation-of-mean-average-precision</p>\r\n<p>Only the rank of your predictions matters.&nbsp; There is no ranking of the actual links in the test set.</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 12017,
      "author_name": "gkoundry",
      "author_url": "",
      "post_date": "06/29/2012 12:12:02",
      "content": "<p>I think you are misunderstanding how the scoring for this competition works.&nbsp; Have a look at this thread: https://www.kaggle.com/c/FacebookRecruiting/forums/t/2002/alternate-explanation-of-mean-average-precision</p>\r\n<p>Only the rank of your predictions matters.&nbsp; There is no ranking of the actual links in the test set.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "12012": "",
    "12017": ""
  },
  "source": "meta"
}