{
  "id": 372139,
  "title": "Imbalanced data",
  "url": "/competitions/2022-ucs757-lab-1-exam-new/discussion/372139",
  "author_name": "Barbagrande007",
  "post_date": "2022-12-14T08:45:56.289000",
  "votes": 5,
  "comment_count": 0,
  "views": 0,
  "content": "<p>The data is very imbalanced. 50% of the label is category 3. Without any over/under sampling or threshold moving you will get an accuracy of 50 percent without the model even \"learning\" the difference between the various label categories. The model will just show a bias for category 3.</p>",
  "messages": [
    {
      "id": 2065033,
      "postDate": "2022-12-14T08:45:56.290Z",
      "content": "<p>The data is very imbalanced. 50% of the label is category 3. Without any over/under sampling or threshold moving you will get an accuracy of 50 percent without the model even \"learning\" the difference between the various label categories. The model will just show a bias for category 3.</p>",
      "rawMarkdown": "The data is very imbalanced. 50% of the label is category 3. Without any over/under sampling or threshold moving you will get an accuracy of 50 percent without the model even \"learning\" the difference between the various label categories. The model will just show a bias for category 3.",
      "votes": 5
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "2065033": "The data is very imbalanced. 50% of the label is category 3. Without any over/under sampling or threshold moving you will get an accuracy of 50 percent without the model even \"learning\" the difference between the various label categories. The model will just show a bias for category 3."
  }
}