{
  "id": 451058,
  "title": "Test set labels",
  "url": "/competitions/child-mind-institute-detect-sleep-states/discussion/451058",
  "author_name": "",
  "post_date": "2023-10-27T01:52:16.618384600Z",
  "votes": 1,
  "comment_count": 3,
  "views": 0,
  "content": "<p>Hi kagglers,<br>\nI have 2 questions even if they may sound stupid.<br>\n1- Does test data have labels, because honestly i can't find them in the train_event dataset??<br>\n2- What does represent \"event\" column in the submission file??<br>\nThanks</p>",
  "messages": [
    {
      "id": "2500749",
      "postDate": "10/27/2023 01:52:16",
      "content": "<p>Hi kagglers,<br>\nI have 2 questions even if they may sound stupid.<br>\n1- Does test data have labels, because honestly i can't find them in the train_event dataset??<br>\n2- What does represent \"event\" column in the submission file??<br>\nThanks</p>",
      "rawMarkdown": "Hi kagglers,\nI have 2 questions even if they may sound stupid.\n1- Does test data have labels, because honestly i can't find them in the train_event dataset??\n2- What does represent \"event\" column in the submission file??\nThanks",
      "votes": null
    },
    {
      "id": "2500794",
      "postDate": "10/27/2023 02:53:00",
      "content": "<ol>\n<li>We have train_series (i.e. raw features) and their corresponding train_events (i.e. labels). We should train our model in order to find a suitable mapping function. Then, we input the test_series data to our model and predict the corresponding test labels (which are not available).</li>\n<li>event refers to type of event which is either 'onset' or 'wakeup'. You can also find this column in train_events.</li>\n</ol>",
      "rawMarkdown": "1. We have train_series (i.e. raw features) and their corresponding train_events (i.e. labels). We should train our model in order to find a suitable mapping function. Then, we input the test_series data to our model and predict the corresponding test labels (which are not available).\n2. event refers to type of event which is either 'onset' or 'wakeup'. You can also find this column in train_events.",
      "votes": null
    },
    {
      "id": "2501494",
      "postDate": "10/27/2023 13:48:24",
      "content": "<p>Hi <a href=\"https://www.kaggle.com/awqatak\" target=\"_blank\">@awqatak</a>, thanks for your reply<br>\nSo from what i understood, test labels are not available in the train_events file, so the \"event\" column in submission represent the predicted labels (correct me if i am wrong).<br>\nThank you so much</p>",
      "rawMarkdown": "Hi @awqatak, thanks for your reply\nSo from what i understood, test labels are not available in the train_events file, so the \"event\" column in submission represent the predicted labels (correct me if i am wrong).\nThank you so much",
      "votes": null
    },
    {
      "id": "2501976",
      "postDate": "10/27/2023 19:58:22",
      "content": "<p>For every series_id, we are predicting the following 'labels': </p>\n<ul>\n<li>the step at which the event occurs.</li>\n<li>the event type</li>\n<li>the confidence score<br>\nkeep in mind that a single series _id could have multiple events.</li>\n</ul>",
      "rawMarkdown": "For every series_id, we are predicting the following 'labels': \n- the step at which the event occurs.\n- the event type\n- the confidence score\nkeep in mind that a single series _id could have multiple events.",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2500794,
      "author_name": "awqatak",
      "author_url": "",
      "post_date": "10/27/2023 02:53:00",
      "content": "<ol>\n<li>We have train_series (i.e. raw features) and their corresponding train_events (i.e. labels). We should train our model in order to find a suitable mapping function. Then, we input the test_series data to our model and predict the corresponding test labels (which are not available).</li>\n<li>event refers to type of event which is either 'onset' or 'wakeup'. You can also find this column in train_events.</li>\n</ol>",
      "votes": null,
      "replies": [
        {
          "id": 2501494,
          "author_name": "anas47",
          "author_url": "",
          "post_date": "10/27/2023 13:48:24",
          "content": "<p>Hi <a href=\"https://www.kaggle.com/awqatak\" target=\"_blank\">@awqatak</a>, thanks for your reply<br>\nSo from what i understood, test labels are not available in the train_events file, so the \"event\" column in submission represent the predicted labels (correct me if i am wrong).<br>\nThank you so much</p>",
          "votes": null,
          "replies": [
            {
              "id": 2501976,
              "author_name": "awqatak",
              "author_url": "",
              "post_date": "10/27/2023 19:58:22",
              "content": "<p>For every series_id, we are predicting the following 'labels': </p>\n<ul>\n<li>the step at which the event occurs.</li>\n<li>the event type</li>\n<li>the confidence score<br>\nkeep in mind that a single series _id could have multiple events.</li>\n</ul>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2500749": "Hi kagglers,\nI have 2 questions even if they may sound stupid.\n1- Does test data have labels, because honestly i can't find them in the train_event dataset??\n2- What does represent \"event\" column in the submission file??\nThanks",
    "2500794": "1. We have train_series (i.e. raw features) and their corresponding train_events (i.e. labels). We should train our model in order to find a suitable mapping function. Then, we input the test_series data to our model and predict the corresponding test labels (which are not available).\n2. event refers to type of event which is either 'onset' or 'wakeup'. You can also find this column in train_events.",
    "2501494": "Hi @awqatak, thanks for your reply\nSo from what i understood, test labels are not available in the train_events file, so the \"event\" column in submission represent the predicted labels (correct me if i am wrong).\nThank you so much",
    "2501976": "For every series_id, we are predicting the following 'labels': \n- the step at which the event occurs.\n- the event type\n- the confidence score\nkeep in mind that a single series _id could have multiple events."
  },
  "source": "meta"
}