{
  "id": 384297,
  "title": "The number of questions in each level_group is fixed/constant?",
  "url": "/competitions/predict-student-performance-from-game-play/discussion/384297",
  "author_name": "Thariq Nugrohotomo",
  "post_date": "2023-02-07T11:16:12.661000",
  "votes": 31,
  "comment_count": 2,
  "views": 0,
  "content": "<p>The time-series test API seems to split an event into 3 prediction phases based on the level_group:</p>\n<table>\n<thead>\n<tr>\n<th>level_group</th>\n<th>number of questions</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>0-4</td>\n<td>3 (q1 until q3)</td>\n</tr>\n<tr>\n<td>5-12</td>\n<td>10 (q4 until q13)</td>\n</tr>\n<tr>\n<td>13-22</td>\n<td>5 (q14 until q18)</td>\n</tr>\n</tbody>\n</table>\n<p>Is this split constant/fixed, or may vary?<br>\nDoes it apply to <code>train_labels.csv</code> as well?</p>\n<p><code>train_labels.csv</code> does not have session_level nor level_group column, hence I'm not sure how to associate it into level_group coming from <code>train.csv</code>.</p>\n<p>Could someone clarify this?<br>\nThanks.</p>",
  "messages": [
    {
      "id": 2133331,
      "postDate": "2023-02-07T11:16:12.660Z",
      "content": "<p>The time-series test API seems to split an event into 3 prediction phases based on the level_group:</p>\n<table>\n<thead>\n<tr>\n<th>level_group</th>\n<th>number of questions</th>\n</tr>\n</thead>\n<tbody>\n<tr>\n<td>0-4</td>\n<td>3 (q1 until q3)</td>\n</tr>\n<tr>\n<td>5-12</td>\n<td>10 (q4 until q13)</td>\n</tr>\n<tr>\n<td>13-22</td>\n<td>5 (q14 until q18)</td>\n</tr>\n</tbody>\n</table>\n<p>Is this split constant/fixed, or may vary?<br>\nDoes it apply to <code>train_labels.csv</code> as well?</p>\n<p><code>train_labels.csv</code> does not have session_level nor level_group column, hence I'm not sure how to associate it into level_group coming from <code>train.csv</code>.</p>\n<p>Could someone clarify this?<br>\nThanks.</p>",
      "rawMarkdown": "The time-series test API seems to split an event into 3 prediction phases based on the level_group:\n\n| level_group | number of questions |\n| --- | --- |\n| 0-4 | 3 (q1 until q3) |\n| 5-12 | 10 (q4 until q13) |\n| 13-22 | 5 (q14 until q18) |\n\nIs this split constant/fixed, or may vary?\nDoes it apply to `train_labels.csv` as well?\n\n`train_labels.csv` does not have session_level nor level_group column, hence I'm not sure how to associate it into level_group coming from `train.csv`.\n\nCould someone clarify this?\nThanks.",
      "votes": 30
    },
    {
      "id": 2133760,
      "postDate": "2023-02-07T15:29:21.953Z",
      "content": "<p>This is my understanding of how the API works too. Regarding <code>train_labels.csv</code> we have the results of all 18 questions and for each user we have level_group 0-4, 5-12, 13-22 data in <code>train.csv</code>. Therefore we can manipulate the train data to match the test data.</p>",
      "rawMarkdown": "This is my understanding of how the API works too. Regarding `train_labels.csv` we have the results of all 18 questions and for each user we have level_group 0-4, 5-12, 13-22 data in `train.csv`. Therefore we can manipulate the train data to match the test data.",
      "votes": 11
    },
    {
      "id": 2133423,
      "postDate": "2023-02-07T12:03:04.610Z",
      "rawMarkdown": "",
      "isDeleted": true
    }
  ],
  "comments": [
    {
      "id": 2133760,
      "author_name": "Chris Deotte",
      "author_url": "",
      "post_date": "2023-02-07T15:29:21.953000",
      "content": "<p>This is my understanding of how the API works too. Regarding <code>train_labels.csv</code> we have the results of all 18 questions and for each user we have level_group 0-4, 5-12, 13-22 data in <code>train.csv</code>. Therefore we can manipulate the train data to match the test data.</p>",
      "votes": 11,
      "replies": []
    },
    {
      "id": 2133423,
      "author_name": "",
      "author_url": "",
      "post_date": "2023-02-07T12:03:04.610000",
      "content": "",
      "votes": 0,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "2133331": "The time-series test API seems to split an event into 3 prediction phases based on the level_group:\n\n| level_group | number of questions |\n| --- | --- |\n| 0-4 | 3 (q1 until q3) |\n| 5-12 | 10 (q4 until q13) |\n| 13-22 | 5 (q14 until q18) |\n\nIs this split constant/fixed, or may vary?\nDoes it apply to `train_labels.csv` as well?\n\n`train_labels.csv` does not have session_level nor level_group column, hence I'm not sure how to associate it into level_group coming from `train.csv`.\n\nCould someone clarify this?\nThanks.",
    "2133760": "This is my understanding of how the API works too. Regarding `train_labels.csv` we have the results of all 18 questions and for each user we have level_group 0-4, 5-12, 13-22 data in `train.csv`. Therefore we can manipulate the train data to match the test data.",
    "2133423": ""
  }
}