Enhancing PCOS Diagnosis with Hybrid ML: Integrating Ultrasound and Clinical Data

This work presents a hybrid machine learning approach that combines clinical data and ultrasound imaging to improve the precision of PCOS diagnosis. With an accuracy of 87% and appropriate for initial screening, the Regular approach uses Logistic Regression on 15 clinical inputs. By contrast, the Hybrid approach achieves a 96% accuracy rate by combining 29 clinical inputs with ultrasound-derived features using a Gradient Boosting Classifier. A Flask backend allows for smooth data processing and real-time predictions, while a web-based interface built with HTML, CSS, and JavaScript makes user interaction easier. The findings demonstrate the efficacy of hybrid techniques in clinical settings and point to their potential for improving diagnostic precision.