{
  "id": 492336,
  "title": "How is test set eeg_id selected? ",
  "url": "/competitions/hms-harmful-brain-activity-classification/discussion/492336",
  "author_name": "",
  "post_date": "2024-04-09T11:00:15.807995200Z",
  "votes": 1,
  "comment_count": 2,
  "views": 0,
  "content": "<p>Is it in a distribution same as our train dataset with only 1 eeg subid?<br>\nOr is it one subid selected from the middle?<br>\nOr is it randomly selected from the eeg sub ids?<br>\nI assumed it is randomly slected from all eeg sub ids, so I evaluated all sub ids and calc mean for each eeg id, but at last week I found cv and LB might not be very well correlated.</p>",
  "messages": [
    {
      "id": "2743290",
      "postDate": "04/09/2024 11:00:15",
      "content": "<p>Is it in a distribution same as our train dataset with only 1 eeg subid?<br>\nOr is it one subid selected from the middle?<br>\nOr is it randomly selected from the eeg sub ids?<br>\nI assumed it is randomly slected from all eeg sub ids, so I evaluated all sub ids and calc mean for each eeg id, but at last week I found cv and LB might not be very well correlated.</p>",
      "rawMarkdown": "Is it in a distribution same as our train dataset with only 1 eeg subid?\nOr is it one subid selected from the middle?\nOr is it randomly selected from the eeg sub ids?\nI assumed it is randomly slected from all eeg sub ids, so I evaluated all sub ids and calc mean for each eeg id, but at last week I found cv and LB might not be very well correlated.",
      "votes": null
    },
    {
      "id": "2743662",
      "postDate": "04/09/2024 15:09:47",
      "content": "<p>We will say a few words in our solution, but I think <a href=\"https://github.com/bdsp-core/IIIC-SPaRCNet/blob/main/IIIC_Classification-Supplemental.pdf\" target=\"_blank\">https://github.com/bdsp-core/IIIC-SPaRCNet/blob/main/IIIC_Classification-Supplemental.pdf</a> explains how host augmented the train data by shifting the eeg data and thats the main reason for multiple rows for same eeg_ids. For test my guess is that they just did not apply this augmentation</p>",
      "rawMarkdown": "We will say a few words in our solution, but I think https://github.com/bdsp-core/IIIC-SPaRCNet/blob/main/IIIC_Classification-Supplemental.pdf explains how host augmented the train data by shifting the eeg data and thats the main reason for multiple rows for same eeg_ids. For test my guess is that they just did not apply this augmentation",
      "votes": null
    },
    {
      "id": "2743717",
      "postDate": "04/09/2024 15:29:22",
      "content": "<p>Cool, thanks Dieter, waiting to see your solution. Congratulations, once again great jump at the last days!</p>",
      "rawMarkdown": "Cool, thanks Dieter, waiting to see your solution. Congratulations, once again great jump at the last days!",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 2743662,
      "author_name": "christofhenkel",
      "author_url": "",
      "post_date": "04/09/2024 15:09:47",
      "content": "<p>We will say a few words in our solution, but I think <a href=\"https://github.com/bdsp-core/IIIC-SPaRCNet/blob/main/IIIC_Classification-Supplemental.pdf\" target=\"_blank\">https://github.com/bdsp-core/IIIC-SPaRCNet/blob/main/IIIC_Classification-Supplemental.pdf</a> explains how host augmented the train data by shifting the eeg data and thats the main reason for multiple rows for same eeg_ids. For test my guess is that they just did not apply this augmentation</p>",
      "votes": null,
      "replies": [
        {
          "id": 2743717,
          "author_name": "goldenlock",
          "author_url": "",
          "post_date": "04/09/2024 15:29:22",
          "content": "<p>Cool, thanks Dieter, waiting to see your solution. Congratulations, once again great jump at the last days!</p>",
          "votes": null,
          "replies": []
        }
      ]
    }
  ],
  "raw_markdown_by_id": {
    "2743290": "Is it in a distribution same as our train dataset with only 1 eeg subid?\nOr is it one subid selected from the middle?\nOr is it randomly selected from the eeg sub ids?\nI assumed it is randomly slected from all eeg sub ids, so I evaluated all sub ids and calc mean for each eeg id, but at last week I found cv and LB might not be very well correlated.",
    "2743662": "We will say a few words in our solution, but I think https://github.com/bdsp-core/IIIC-SPaRCNet/blob/main/IIIC_Classification-Supplemental.pdf explains how host augmented the train data by shifting the eeg data and thats the main reason for multiple rows for same eeg_ids. For test my guess is that they just did not apply this augmentation",
    "2743717": "Cool, thanks Dieter, waiting to see your solution. Congratulations, once again great jump at the last days!"
  },
  "source": "meta"
}