Leveraging Machine Learning for Early Detection and Intervention in Polycystic Ovary Syndrome (PCOS): A Predictive Model Approach

The present world is concerned with every disorder. Polycystic Ovary Syndrome (PCOS) is one of the most common endocrine disorders. It affects women of reproductive age, with implications for psychological and metabolic health. So, it is an exigence for accurate diagnosis of PCOS for effective treatment and management for women. It also remains a contravention to predict the syndrome by some particular symptoms. This paper explores the application of machine learning techniques to predict whether a woman is at risk of PCOS (labeled as “sick”) or not (“normal”). We preprocess the dataset by handling null values, encoding nonnumerical features, and balancing the data using the Synthetic Minority Over-sampling Technique (SMOTE) to relieve class imbalance. Classification models are trained using machine learning techniques, such as Random Forest, Gradient Boosting, Gaussian Naive Bias, Decision Tree, and Logistic Regression. The maximum accuracy gain is 96.89% for the Gradient boosting classifier. Also compare the value of recall, precision, f1-score, and kappa score among the classifiers. The findings of this research in the application of machine learning accentuate its potential for early detection and intervention of PCOS.