An Advanced Ensemble Machine Learning Framework for Early Prediction of Pcos With Real-Time Clinical Decision Support

Abstract- Polycystic Ovary Syndrome is a prevalent and multifaceted endocrine disorder affecting women of reproductive age, often leading to reproductive, metabolic, and psychological complications. Early diagnosis remains a significant challenge due to heterogeneous symptoms, dependence on clinical expertise, and the need for multiple diagnostic procedures such as hormonal analysis and ultrasound imaging. These limitations highlight the necessity for an efficient, cost-effective, and automated predictive system. This study proposes an advanced ensemble machine learning framework for the early prediction of PCOS using clinical and symptomatic data. The proposed approach integrates robust data preprocessing techniques, including missing value imputation, feature scaling, and Synthetic Minority Oversampling Technique (SMOTE) to address class imbalance. Furthermore, Recursive Feature Elimination (RFE) is employed to identify the most significant predictive features, thereby enhancing model performance and reducing computational complexity. The predictive model combines two powerful ensemble algorithms—Random Forest and Extreme Gradient Boosting (XGBoost)—to leverage their complementary strengths in handling nonlinear relationships and improving classification accuracy. The model is trained and evaluated on a publicly available dataset, achieving an accuracy of 97% and a Receiver Operating Characteristic Area Under Curve (ROC-AUC) score of 0.99, indicating excellent predictive capability. In addition to model development, a real-time web-based application is implemented to facilitate user interaction and practical deployment in healthcare environments. The system enables users to input clinical parameters and receive instant predictions, thereby supporting early diagnosis and clinical decision-making. The experimental results demonstrate that the proposed framework significantly outperforms traditional machine learning models and offers a scalable solution for PCOS prediction. This research contributes to the advancement of intelligent healthcare systems by providing a reliable, accurate, and accessible tool for early detection, which can ultimately improve patient outcomes and reduce the burden on healthcare systems.