{
  "id": 145943,
  "title": "Use simple mask features to predict",
  "url": "/competitions/prostate-cancer-grade-assessment/discussion/145943",
  "author_name": "hawkey",
  "post_date": "2020-04-25T06:37:54.626000",
  "votes": 1,
  "comment_count": 0,
  "views": 0,
  "content": "<p>I tried to use simple percent features extracted from training masks to predict isup_grade. The idea is from <a href=\"https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/145144\">this</a> discussion. \nThe results on radboud training data is promising, and this maybe a startpoint for segmentation models.</p>\n\n<p>Results: <a href=\"https://www.kaggle.com/hawkey/mask-only-simple-classifiers-cv-0-81\">https://www.kaggle.com/hawkey/mask-only-simple-classifiers-cv-0-81</a></p>",
  "messages": [
    {
      "id": 820119,
      "postDate": "2020-04-25T06:37:54.627Z",
      "content": "<p>I tried to use simple percent features extracted from training masks to predict isup_grade. The idea is from <a href=\"https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/145144\">this</a> discussion. \nThe results on radboud training data is promising, and this maybe a startpoint for segmentation models.</p>\n\n<p>Results: <a href=\"https://www.kaggle.com/hawkey/mask-only-simple-classifiers-cv-0-81\">https://www.kaggle.com/hawkey/mask-only-simple-classifiers-cv-0-81</a></p>",
      "rawMarkdown": "I tried to use simple percent features extracted from training masks to predict isup_grade. The idea is from [this](https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/145144) discussion. \nThe results on radboud training data is promising, and this maybe a startpoint for segmentation models.\n\nResults: https://www.kaggle.com/hawkey/mask-only-simple-classifiers-cv-0-81",
      "votes": 1
    }
  ],
  "comments": [],
  "raw_markdown_by_id": {
    "820119": "I tried to use simple percent features extracted from training masks to predict isup_grade. The idea is from [this](https://www.kaggle.com/c/prostate-cancer-grade-assessment/discussion/145144) discussion. \nThe results on radboud training data is promising, and this maybe a startpoint for segmentation models.\n\nResults: https://www.kaggle.com/hawkey/mask-only-simple-classifiers-cv-0-81"
  }
}