{
  "id": 21694,
  "title": "Test set unseen labels?",
  "url": "/competitions/facebook-v-predicting-check-ins/discussion/21694",
  "author_name": "",
  "post_date": "2016-06-15T07:44:34.987Z",
  "votes": 2,
  "comment_count": 2,
  "views": 376,
  "content": "<p>@admins, the data page says: 'The train and test dataset are split based on time', but then is possible there are new labels in test set. \nDid you clean this new labels in test set or they are there, like one more part of the problem? (obviously they are impossible to predict).</p>",
  "messages": [
    {
      "id": "124051",
      "postDate": "06/15/2016 07:44:34",
      "content": "<p>@admins, the data page says: 'The train and test dataset are split based on time', but then is possible there are new labels in test set. \nDid you clean this new labels in test set or they are there, like one more part of the problem? (obviously they are impossible to predict).</p>",
      "rawMarkdown": "admins, the data page says: 'The train and test dataset are split based on time', but then is possible there are new labels in test set. \r\nDid you clean this new labels in test set or they are there, like one more part of the problem? (obviously they are impossible to predict).",
      "votes": null
    },
    {
      "id": "124119",
      "postDate": "06/15/2016 17:18:18",
      "content": "<p>@Jose, </p>\n\n<p>Good question. We did remove the unseen businesses from the test set because of the reason you stated. </p>",
      "rawMarkdown": "Jose, \r\n\r\nGood question. We did remove the unseen businesses from the test set because of the reason you stated.",
      "votes": null
    },
    {
      "id": "124122",
      "postDate": "06/15/2016 17:23:10",
      "content": "<p>[quote=Wendy Kan;124119]</p>\n\n<p>@Jose, </p>\n\n<p>Good question. We did remove the unseen businesses from the test set because of the reason you stated. </p>\n\n<p>[/quote]</p>\n\n<p>Thank you, Wendy. This explains the better behavior of the lb score vs using time split cv (without clean the unseen place_ids)</p>",
      "rawMarkdown": "[quote=Wendy Kan;124119]\r\n\r\n@Jose, \r\n\r\nGood question. We did remove the unseen businesses from the test set because of the reason you stated. \r\n\r\n[/quote]\r\n\r\nThank you, Wendy. This explains the better behavior of the lb score vs using time split cv (without clean the unseen place_ids)",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 124119,
      "author_name": "wendykan",
      "author_url": "",
      "post_date": "06/15/2016 17:18:18",
      "content": "<p>@Jose, </p>\n\n<p>Good question. We did remove the unseen businesses from the test set because of the reason you stated. </p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 124122,
      "author_name": "blindape",
      "author_url": "",
      "post_date": "06/15/2016 17:23:10",
      "content": "<p>[quote=Wendy Kan;124119]</p>\n\n<p>@Jose, </p>\n\n<p>Good question. We did remove the unseen businesses from the test set because of the reason you stated. </p>\n\n<p>[/quote]</p>\n\n<p>Thank you, Wendy. This explains the better behavior of the lb score vs using time split cv (without clean the unseen place_ids)</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "124051": "admins, the data page says: 'The train and test dataset are split based on time', but then is possible there are new labels in test set. \r\nDid you clean this new labels in test set or they are there, like one more part of the problem? (obviously they are impossible to predict).",
    "124119": "Jose, \r\n\r\nGood question. We did remove the unseen businesses from the test set because of the reason you stated.",
    "124122": "[quote=Wendy Kan;124119]\r\n\r\n@Jose, \r\n\r\nGood question. We did remove the unseen businesses from the test set because of the reason you stated. \r\n\r\n[/quote]\r\n\r\nThank you, Wendy. This explains the better behavior of the lb score vs using time split cv (without clean the unseen place_ids)"
  },
  "source": "meta"
}