{
  "id": 27635,
  "title": "Basic question regarding final evaluation",
  "url": "/competitions/outbrain-click-prediction/discussion/27635",
  "author_name": "",
  "post_date": "2017-01-12T09:16:44.297Z",
  "votes": null,
  "comment_count": 2,
  "views": 106,
  "content": "<p>On \"my submission\" page, it is written: \"Your final score will not be based on the same exact subset data as the public leaderboard, but rather a different private data subset of your full submission\"</p>\n\n<p>I am under the assumption that we just need to submit the output as we predict on clicks_test.csv file provided. Now, if that is true then final evaluation has to be on the subset of rows present in clicks_test.csv. So, then does the above line simply mean that public leaderboard score is evaluated on a subset of clicks_test.csv and private leaderboard would be evaluated may be evaluated on another subset?</p>\n\n<p>Is my interpretation correct?</p>",
  "messages": [
    {
      "id": "155654",
      "postDate": "01/12/2017 09:16:44",
      "content": "<p>On \"my submission\" page, it is written: \"Your final score will not be based on the same exact subset data as the public leaderboard, but rather a different private data subset of your full submission\"</p>\n\n<p>I am under the assumption that we just need to submit the output as we predict on clicks_test.csv file provided. Now, if that is true then final evaluation has to be on the subset of rows present in clicks_test.csv. So, then does the above line simply mean that public leaderboard score is evaluated on a subset of clicks_test.csv and private leaderboard would be evaluated may be evaluated on another subset?</p>\n\n<p>Is my interpretation correct?</p>",
      "rawMarkdown": "On \"my submission\" page, it is written: \"Your final score will not be based on the same exact subset data as the public leaderboard, but rather a different private data subset of your full submission\"\r\n\r\nI am under the assumption that we just need to submit the output as we predict on clicks_test.csv file provided. Now, if that is true then final evaluation has to be on the subset of rows present in clicks_test.csv. So, then does the above line simply mean that public leaderboard score is evaluated on a subset of clicks_test.csv and private leaderboard would be evaluated may be evaluated on another subset?\r\n\r\nIs my interpretation correct?",
      "votes": null
    },
    {
      "id": "155658",
      "postDate": "01/12/2017 09:33:05",
      "content": "<p>Yes, that's correct. Go to the leaderboard page and you'll see the following statement:\n\"<em>This leaderboard is calculated on approximately 30% of the test data.\nThe final results will be based on the other 70%, so the final standings may be different.</em>\"</p>",
      "rawMarkdown": "Yes, that's correct. Go to the leaderboard page and you'll see the following statement:\r\n\"*This leaderboard is calculated on approximately 30% of the test data.\r\nThe final results will be based on the other 70%, so the final standings may be different.*\"",
      "votes": null
    },
    {
      "id": "155660",
      "postDate": "01/12/2017 09:37:17",
      "content": "<p>Thanks kos_</p>",
      "rawMarkdown": "Thanks kos_",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 155658,
      "author_name": "phansoks",
      "author_url": "",
      "post_date": "01/12/2017 09:33:05",
      "content": "<p>Yes, that's correct. Go to the leaderboard page and you'll see the following statement:\n\"<em>This leaderboard is calculated on approximately 30% of the test data.\nThe final results will be based on the other 70%, so the final standings may be different.</em>\"</p>",
      "votes": null,
      "replies": []
    },
    {
      "id": 155660,
      "author_name": "aknirala",
      "author_url": "",
      "post_date": "01/12/2017 09:37:17",
      "content": "<p>Thanks kos_</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "155654": "On \"my submission\" page, it is written: \"Your final score will not be based on the same exact subset data as the public leaderboard, but rather a different private data subset of your full submission\"\r\n\r\nI am under the assumption that we just need to submit the output as we predict on clicks_test.csv file provided. Now, if that is true then final evaluation has to be on the subset of rows present in clicks_test.csv. So, then does the above line simply mean that public leaderboard score is evaluated on a subset of clicks_test.csv and private leaderboard would be evaluated may be evaluated on another subset?\r\n\r\nIs my interpretation correct?",
    "155658": "Yes, that's correct. Go to the leaderboard page and you'll see the following statement:\r\n\"*This leaderboard is calculated on approximately 30% of the test data.\r\nThe final results will be based on the other 70%, so the final standings may be different.*\"",
    "155660": "Thanks kos_"
  },
  "source": "meta"
}