{
  "id": 416018,
  "title": "Question about session_id, timeframe and test data",
  "url": "/competitions/predict-student-performance-from-game-play/discussion/416018",
  "author_name": "Peter Chuikov",
  "post_date": "2023-06-09T07:08:07.399000",
  "votes": 0,
  "comment_count": 0,
  "views": 0,
  "content": "<p>Hi everyone!</p>\n<p>I see, that we can split session_id into time, so we can using time series cross-validation and in general to deal with data as time series. </p>\n<p>But I wonder, the test data is shift to the future from current data, isn't it? <br>\nCV strategy must depends on that information.</p>",
  "messages": [
    {
      "id": 2293421,
      "postDate": "2023-06-09T07:08:07.400Z",
      "content": "<p>Hi everyone!</p>\n<p>I see, that we can split session_id into time, so we can using time series cross-validation and in general to deal with data as time series. </p>\n<p>But I wonder, the test data is shift to the future from current data, isn't it? <br>\nCV strategy must depends on that information.</p>",
      "rawMarkdown": "Hi everyone!\n\nI see, that we can split session_id into time, so we can using time series cross-validation and in general to deal with data as time series. \n\nBut I wonder, the test data is shift to the future from current data, isn't it? \nCV strategy must depends on that information."
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2293421": "Hi everyone!\n\nI see, that we can split session_id into time, so we can using time series cross-validation and in general to deal with data as time series. \n\nBut I wonder, the test data is shift to the future from current data, isn't it? \nCV strategy must depends on that information."
  }
}