A Comprehensive Review of Predicting Lifestyle-Based Disease Specifically PCOS Among Women Using Data Mining and Machine Learning Approaches

Urbanization and industrialization are increasing risk factors associated with lifestyle choices, contributing to the advancement in lifestyle-related diseases. This primarily results from habits that encourage a sedentary lifestyle, leading to various health issues and ultimately peaking in chronic Non-Communicable Diseases (NCDs). Numerous lifestyle-based diseases affecting women globally include cardiovascular disease, breast cancer, PCOS, type-2 diabetes, thyroid disorders, and more. PCOS, or polycystic ovary syndrome, stands out as a prevalent health concern affecting a significant proportion of women worldwide. The timely identification and accurate prediction of PCOS are crucial. Numerous researchers have dedicated their efforts to predicting PCOS using data mining and machine learning approaches. This Comprehensive Review (CR) seeks to thoroughly examine and summarize various disease prediction models proposed in current research for predicting PCOS. The paper scrutinizes technical aspects such as data collection selection, feature engineering applications, data preprocessing, characteristics of data mining and machine learning techniques, methodology, and evaluation matrices. However, none of the selected studies conducted external validation. The paper also explores the limitations and outlines the scope for future research perspectives in predicting PCOS, presenting opportunities for scholars interested in exploring additional methods to anticipate early-stage risks associated with PCOS.