Scholarlink

The Polycystic Ovary Syndrome (PCOS) is one of the most common endocrine disorder; affecting up to 13% of women of reproductive age. However up to 70% of cases go undiagnosed. Machine learning (ML) helps with detection but raises privacy concerns, since we’re dealing with sensitive patients’ information. Federated Learning (FL) serves as privacy-preserving alternative, since it allows training models across institutions without sharing data. This study compares centralized (88.7% accuracy) and federated (71.56% accuracy) learning for PCOS diagnosis. Although, FL ensures privacy, its performance is hindered by data inconsistency and non-IID distribution. Future improvements in FL could bridge this gap, making it a viable approach for secure healthcare AI.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

PCOS Diagnosis: Comparison of Centralized and Federated Learning

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Aminetou Tekrour,
  • Moulay Youssef Hadi

摘要

The Polycystic Ovary Syndrome (PCOS) is one of the most common endocrine disorder; affecting up to 13% of women of reproductive age. However up to 70% of cases go undiagnosed. Machine learning (ML) helps with detection but raises privacy concerns, since we’re dealing with sensitive patients’ information. Federated Learning (FL) serves as privacy-preserving alternative, since it allows training models across institutions without sharing data. This study compares centralized (88.7% accuracy) and federated (71.56% accuracy) learning for PCOS diagnosis. Although, FL ensures privacy, its performance is hindered by data inconsistency and non-IID distribution. Future improvements in FL could bridge this gap, making it a viable approach for secure healthcare AI.