Diagnosis of Polycystic Ovary Syndrome Using Machine Learning Models: A Comparative Analysis

Polycystic Ovary Syndrome (PCOS) is a serious health problem affecting a significant percentage of female population worldwide during their reproductive years and adolescence. Besides being one of the leading causes of female infertility, PCOS also carries risks of long-term health issues such as diabetes, high blood pressure, and cardiovascular disease. While PCOS is not curable, proper treatment can alleviate some symptoms. Hence, timely diagnosis is crucial for effective treatment. PCOS diagnosis is challenging due to varying symptoms among affected women. Recently, promising progress in medical diagnosis has been made using machine learning techniques. This paper presents a comparative analysis of the performance of single classifier-based machine learning models and ensemble machine learning models in diagnosing PCOS using benchmark Kaggle PCOS dataset. 10-fold cross-validation has been used to assess the performance of the models. From the experimental results it is observed that among the thirteen different machine learning models considered for this study, the Random Forest classifier has achieved the best performance with an accuracy and F1-score of 90.1% and 83.6% respectively.