{
  "id": 2578,
  "title": "Groups in test",
  "url": "/competitions/Raising-Money-to-Fund-an-Organizational-Mission/discussion/2578",
  "author_name": "",
  "post_date": "2012-09-05T09:17:22.333Z",
  "votes": null,
  "comment_count": 1,
  "views": 1562,
  "content": "<p>Hi there,</p>\r\n<p>I wonder, where does the groups in the test set actually come from? Is it just a combined key of keys already in the datasets? I'm asking, because if you e.g. want to cross-validate different models using the training set, there are is no explicit grouping\r\n available. So how is the grouping in the test set actually done?</p>\r\n<p>Thanks in advance for your help.</p>",
  "messages": [
    {
      "id": "13934",
      "postDate": "09/05/2012 09:17:22",
      "content": "<p>Hi there,</p>\r\n<p>I wonder, where does the groups in the test set actually come from? Is it just a combined key of keys already in the datasets? I'm asking, because if you e.g. want to cross-validate different models using the training set, there are is no explicit grouping\r\n available. So how is the grouping in the test set actually done?</p>\r\n<p>Thanks in advance for your help.</p>",
      "rawMarkdown": "",
      "votes": null
    },
    {
      "id": "13950",
      "postDate": "09/05/2012 15:23:40",
      "content": "<p>The grouping for the Evaluation dataset is by packageid and phase. This combination represents a single mailing, and we want to identify the top 75% of each one of these mailings.\r\n</p>\r\n<p>Thanks!</p>",
      "rawMarkdown": "",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 13950,
      "author_name": "fundraising",
      "author_url": "",
      "post_date": "09/05/2012 15:23:40",
      "content": "<p>The grouping for the Evaluation dataset is by packageid and phase. This combination represents a single mailing, and we want to identify the top 75% of each one of these mailings.\r\n</p>\r\n<p>Thanks!</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "13934": "",
    "13950": ""
  },
  "source": "meta"
}