Automated screening of Polycystic Ovary Syndrome using machine learning techniques

Abstract

Polycystic Ovary Syndrome (PCOS) is one of the most common type of endocrine disorder in reproductive age women. This may result in infertility and anovulation. The diagnostic criterion includes the clinical and metabolic parameters which act as an early marker for the disease. We described a method that automates the PCOS detection based on these markers. Our algorithm involves the formulation of feature vector based on the clinical and metabolic features and statistically significant features for discriminating between normal and PCOS groups are selected based on two sample t-test. To classify the selected feature Bayesian and Logistic Regression (LR) classifier are used. An automated system will act as an assisted tool for the doctor for saving considerable time in examining the patients and hence reducing the delay in diagnosing the risk of PCOS. The study demonstrated that the performance of Bayesian classifier is better than the logistic regression. The overall accuracy of Bayesian classifier is 93.93% as compared with logistic regression i.e. 91.04%.