{
  "id": 10056,
  "title": "Ratio of Seizures to Nonseizures in Test Data",
  "url": "/competitions/seizure-detection/discussion/10056",
  "author_name": "",
  "post_date": "2014-08-16T12:14:23.893Z",
  "votes": null,
  "comment_count": 1,
  "views": 839,
  "content": "<p>Could we have this information: What is the (approximate) ratio of seizures to nonseizures in the test data, overall? Like a low ratio ~10%? Or is it balanced like ~50%? For some reason my classification results come heavily biased against deciding for seizures so I am getting few number of seizures in the classification.&nbsp;</p>\n\n<p>Thanks</p>",
  "messages": [
    {
      "id": "52083",
      "postDate": "08/16/2014 12:14:23",
      "content": "<p>Could we have this information: What is the (approximate) ratio of seizures to nonseizures in the test data, overall? Like a low ratio ~10%? Or is it balanced like ~50%? For some reason my classification results come heavily biased against deciding for seizures so I am getting few number of seizures in the classification.&nbsp;</p>\n\n<p>Thanks</p>",
      "rawMarkdown": "Could we have this information: What is the (approximate) ratio of seizures to nonseizures in the test data, overall? Like a low ratio ~10%? Or is it balanced like ~50%? For some reason my classification results come heavily biased against deciding for seizures so I am getting few number of seizures in the classification.&nbsp;\n\nThanks",
      "votes": null
    },
    {
      "id": "3080968",
      "postDate": "12/26/2024 03:05:20",
      "content": "<p>我有个疑问，这个数据集的测试集都没有标签，怎么验证准确性啊，我想用这个，可以不可以把实验数据去掉，留下有标签数据，划分训练和测试集，我只想验证我的算法的有效性，不想参加比赛</p>",
      "rawMarkdown": "我有个疑问，这个数据集的测试集都没有标签，怎么验证准确性啊，我想用这个，可以不可以把实验数据去掉，留下有标签数据，划分训练和测试集，我只想验证我的算法的有效性，不想参加比赛",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 3080968,
      "author_name": "sunyongxin711",
      "author_url": "",
      "post_date": "12/26/2024 03:05:20",
      "content": "<p>我有个疑问，这个数据集的测试集都没有标签，怎么验证准确性啊，我想用这个，可以不可以把实验数据去掉，留下有标签数据，划分训练和测试集，我只想验证我的算法的有效性，不想参加比赛</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "52083": "Could we have this information: What is the (approximate) ratio of seizures to nonseizures in the test data, overall? Like a low ratio ~10%? Or is it balanced like ~50%? For some reason my classification results come heavily biased against deciding for seizures so I am getting few number of seizures in the classification.&nbsp;\n\nThanks",
    "3080968": "我有个疑问，这个数据集的测试集都没有标签，怎么验证准确性啊，我想用这个，可以不可以把实验数据去掉，留下有标签数据，划分训练和测试集，我只想验证我的算法的有效性，不想参加比赛"
  },
  "source": "meta"
}