{
  "id": 133610,
  "title": "Reason for low scores in Notebooks here",
  "url": "/competitions/google-cloud-ncaa-march-madness-2020-division-1-mens-tournament/discussion/133610",
  "author_name": "",
  "post_date": "2020-03-03T13:47:07.205399500Z",
  "votes": 1,
  "comment_count": 1,
  "views": 0,
  "content": "<p>Hi all, this's my 1st competition.  Happy to be here with many Notebooks from members here for help!</p>\n\n<p>I scanned through notebooks here and note that they have a very very low score of below 0.4 (Experts here said this means info leakage). I realised that it's because the training dataset (MNCAATourneyCompactResults) contains 2015 - 2019, where our test dataset requires us to predict 2015 - 2019, hence the low score.</p>\n\n<p>To prevent info leakage here, remove the training dataset from 2015 - 2019 in (MNCAATourneyCompactResults). More specficially remove Day 136 onwards for 2015 \n and all data from 2016 - 2019, since Day 134/135 for 2015 is the pre-selection days.</p>\n\n<p>Thanks for reading and do upvote if it helps!</p>",
  "messages": [
    {
      "id": "762402",
      "postDate": "03/03/2020 13:47:07",
      "content": "<p>Hi all, this's my 1st competition.  Happy to be here with many Notebooks from members here for help!</p>\n\n<p>I scanned through notebooks here and note that they have a very very low score of below 0.4 (Experts here said this means info leakage). I realised that it's because the training dataset (MNCAATourneyCompactResults) contains 2015 - 2019, where our test dataset requires us to predict 2015 - 2019, hence the low score.</p>\n\n<p>To prevent info leakage here, remove the training dataset from 2015 - 2019 in (MNCAATourneyCompactResults). More specficially remove Day 136 onwards for 2015 \n and all data from 2016 - 2019, since Day 134/135 for 2015 is the pre-selection days.</p>\n\n<p>Thanks for reading and do upvote if it helps!</p>",
      "rawMarkdown": "Hi all, this's my 1st competition.  Happy to be here with many Notebooks from members here for help!\n\nI scanned through notebooks here and note that they have a very very low score of below 0.4 (Experts here said this means info leakage). I realised that it's because the training dataset (MNCAATourneyCompactResults) contains 2015 - 2019, where our test dataset requires us to predict 2015 - 2019, hence the low score.\n\nTo prevent info leakage here, remove the training dataset from 2015 - 2019 in (MNCAATourneyCompactResults). More specficially remove Day 136 onwards for 2015 \n and all data from 2016 - 2019, since Day 134/135 for 2015 is the pre-selection days.\n\nThanks for reading and do upvote if it helps!",
      "votes": null
    },
    {
      "id": "763567",
      "postDate": "03/04/2020 15:56:38",
      "content": "<p>People do it, I did it once and couldn't really erase it. It's fine. It'll clear itself.</p>",
      "rawMarkdown": "People do it, I did it once and couldn't really erase it. It's fine. It'll clear itself.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 763567,
      "author_name": "verracodeguacas",
      "author_url": "",
      "post_date": "03/04/2020 15:56:38",
      "content": "<p>People do it, I did it once and couldn't really erase it. It's fine. It'll clear itself.</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "762402": "Hi all, this's my 1st competition.  Happy to be here with many Notebooks from members here for help!\n\nI scanned through notebooks here and note that they have a very very low score of below 0.4 (Experts here said this means info leakage). I realised that it's because the training dataset (MNCAATourneyCompactResults) contains 2015 - 2019, where our test dataset requires us to predict 2015 - 2019, hence the low score.\n\nTo prevent info leakage here, remove the training dataset from 2015 - 2019 in (MNCAATourneyCompactResults). More specficially remove Day 136 onwards for 2015 \n and all data from 2016 - 2019, since Day 134/135 for 2015 is the pre-selection days.\n\nThanks for reading and do upvote if it helps!",
    "763567": "People do it, I did it once and couldn't really erase it. It's fine. It'll clear itself."
  },
  "source": "meta"
}