{
  "id": 403735,
  "title": "Missing labels?",
  "url": "/competitions/predict-student-performance-from-game-play/discussion/403735",
  "author_name": "",
  "post_date": "2023-04-24T15:40:13.823053100Z",
  "votes": 1,
  "comment_count": 4,
  "views": 0,
  "content": "<p>In the data tab, \"At each checkpoint, you will have access to all previous test data for that section.\" implies that our predictive model will have access to the labels for Q 1-3 in order to predict Q 4-13. However, the test.csv includes no such labels. </p>\n<p>I've trained my models to use this data, and several engineered features, from the train and train labels .csv files. Should we have access to the labels for the test.csv session data?</p>",
  "messages": [
    {
      "id": "2232768",
      "postDate": "04/24/2023 15:40:13",
      "content": "<p>In the data tab, \"At each checkpoint, you will have access to all previous test data for that section.\" implies that our predictive model will have access to the labels for Q 1-3 in order to predict Q 4-13. However, the test.csv includes no such labels. </p>\n<p>I've trained my models to use this data, and several engineered features, from the train and train labels .csv files. Should we have access to the labels for the test.csv session data?</p>",
      "rawMarkdown": "In the data tab, \"At each checkpoint, you will have access to all previous test data for that section.\" implies that our predictive model will have access to the labels for Q 1-3 in order to predict Q 4-13. However, the test.csv includes no such labels. \n\nI've trained my models to use this data, and several engineered features, from the train and train labels .csv files. Should we have access to the labels for the test.csv session data?",
      "votes": null
    },
    {
      "id": "2237389",
      "postDate": "04/27/2023 15:38:30",
      "content": "<p>The label for train set is in <code>train_label</code> file.</p>",
      "rawMarkdown": "The label for train set is in `train_label` file.",
      "votes": null
    },
    {
      "id": "2237409",
      "postDate": "04/27/2023 15:52:46",
      "content": "<p>That wasn't my question. For predictions, how will we have access to performance on the first checkpoint data to predict the second? We have it for training, but not for test.</p>\n<p>Specifically, in test, I should have labels for 1, 2, and 3 when I'm predicting 4 - 13. But access to that isn't explained in the data.</p>",
      "rawMarkdown": "That wasn't my question. For predictions, how will we have access to performance on the first checkpoint data to predict the second? We have it for training, but not for test.\n\nSpecifically, in test, I should have labels for 1, 2, and 3 when I'm predicting 4 - 13. But access to that isn't explained in the data.",
      "votes": null
    },
    {
      "id": "2253932",
      "postDate": "05/10/2023 14:20:23",
      "content": "<p>I'm specifically talking about the test data. We have to have a way to receive the student performance on Q 1-3 to predict performance on Q 4-13, and the performance of Q 4-13 to predict Q 14-18. If that performance is a feature of the model, we need to know how to capture it in our code for the predictive piece.</p>",
      "rawMarkdown": "I'm specifically talking about the test data. We have to have a way to receive the student performance on Q 1-3 to predict performance on Q 4-13, and the performance of Q 4-13 to predict Q 14-18. If that performance is a feature of the model, we need to know how to capture it in our code for the predictive piece.",
      "votes": null
    },
    {
      "id": "2253939",
      "postDate": "05/10/2023 14:28:35",
      "content": "<p>Well… You don't. 👀</p>",
      "rawMarkdown": "Well... You don't. 👀",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2237389,
      "author_name": "thaweewatboy",
      "author_url": "",
      "post_date": "04/27/2023 15:38:30",
      "content": "<p>The label for train set is in <code>train_label</code> file.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2253932,
          "author_name": "vbradculbertson",
          "author_url": "",
          "post_date": "05/10/2023 14:20:23",
          "content": "<p>I'm specifically talking about the test data. We have to have a way to receive the student performance on Q 1-3 to predict performance on Q 4-13, and the performance of Q 4-13 to predict Q 14-18. If that performance is a feature of the model, we need to know how to capture it in our code for the predictive piece.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 2237409,
      "author_name": "vbradculbertson",
      "author_url": "",
      "post_date": "04/27/2023 15:52:46",
      "content": "<p>That wasn't my question. For predictions, how will we have access to performance on the first checkpoint data to predict the second? We have it for training, but not for test.</p>\n<p>Specifically, in test, I should have labels for 1, 2, and 3 when I'm predicting 4 - 13. But access to that isn't explained in the data.</p>",
      "votes": null,
      "replies": [
        {
          "id": 2253939,
          "author_name": "woprime",
          "author_url": "",
          "post_date": "05/10/2023 14:28:35",
          "content": "<p>Well… You don't. 👀</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2232768": "In the data tab, \"At each checkpoint, you will have access to all previous test data for that section.\" implies that our predictive model will have access to the labels for Q 1-3 in order to predict Q 4-13. However, the test.csv includes no such labels. \n\nI've trained my models to use this data, and several engineered features, from the train and train labels .csv files. Should we have access to the labels for the test.csv session data?",
    "2237389": "The label for train set is in `train_label` file.",
    "2237409": "That wasn't my question. For predictions, how will we have access to performance on the first checkpoint data to predict the second? We have it for training, but not for test.\n\nSpecifically, in test, I should have labels for 1, 2, and 3 when I'm predicting 4 - 13. But access to that isn't explained in the data.",
    "2253932": "I'm specifically talking about the test data. We have to have a way to receive the student performance on Q 1-3 to predict performance on Q 4-13, and the performance of Q 4-13 to predict Q 14-18. If that performance is a feature of the model, we need to know how to capture it in our code for the predictive piece.",
    "2253939": "Well... You don't. 👀"
  },
  "source": "meta"
}