Comprehensive Review of Polycystic Ovary Syndrome Detection Techniques

Abstract

Polycystic ovary syndrome (PCOS) is a serious hormonal disorder that affects women and significantly impacts their quality of life. In modern times, women are increasingly susceptible to this syndrome, which is a major cause of numerous health problems, most notably infertility. Early detection of PCOS significantly reduces complications, making an early and accurate diagnosis system crucial.

Among all diagnostic techniques, machine learning (ML) has demonstrated superior performance due to its ability to extract features and patterns from data. Therefore, this field has received widespread attention from researchers, and numerous studies have been conducted to detect PCOS using machine learning techniques. These methods have included convolutional neural networks (CNN), support vector machines (SVM), k-nearest neighbors (KNN), random forests, logistic regression, decision trees, and the Naive Bayes algorithm, among others.

This paper aims to shed light on all current techniques used in PCOS detection using machine learning algorithms, providing a comprehensive descriptive and contextual review. It also provides a detailed analysis of how various ML techniques have been used in this field over the past decades, with an in-depth discussion of these approaches. A comprehensive review of the various datasets used in PCOS diagnosis is also provided, comparing the performance of algorithms from both quantitative and qualitative perspectives. Finally, the paper discusses the most prominent challenges facing this field, in addition to exploring future research prospects.