A Comparative Study of Machine Learning Algorithms for PCOS Diagnosis and Prediction

This comparative study examined various machine learning algorithms and image-processing techniques used for the diagnosis and prediction of Polycystic Ovary Syndrome (PCOS). The reviewed literature demonstrates that machine learning models can effectively assist healthcare professionals in identifying PCOS at an early stage using clinical, hormonal, metabolic, and ultrasound imaging data. Among the analyzed algorithms, Random Forest consistently showed superior performance due to its ensemble learning capability, robustness against overfitting, and ability to handle complex nonlinear relationships among clinical features. Support Vector Machine and Logistic Regression also provided reliable classification results, while Decision Tree, Naïve Bayes, and K-Nearest Neighbor offered simpler and interpretable solutions for specific datasets. Furthermore, ultrasound image-based approaches have emerged as promising tools for automated PCOS diagnosis. Segmentation techniques such as Otsu Thresholding and Chan–Vese methods improve follicle detection and ovarian morphology analysis, thereby enhancing diagnostic accuracy. Hybrid systems that combine machine learning with image-processing techniques achieve better performance than standalone approaches. Despite these advancements, challenges such as limited dataset availability, class imbalance, image noise, and lack of model interpretability remain significant barriers to widespread clinical adoption. Therefore, further research is required to develop reliable, explainable, and clinically deployable PCOS diagnostic systems.

[1] A. S. Prapty and T. T. Shitu, ICCIT, 2020. [2] M. Madhumitha et al., IEEE Conference, 2021. [3] N. Nazarudin et al., IEEE Conference, 2020. [4] Rotterdam ESHRE/ASRM Consensus, Human Reproduction, 2004. [5] J. Azziz et al., Nature Reviews Disease Primers, 2016. [6] H. Teede et al., International PCOS Guidelines, 2018. [7] M. O. Goodarzi et al., Endocrine Reviews, 2011. [8] H. F. Escobar-Morreale, Nature Reviews Endocrinology, 2018. [9] E. Diamanti-Kandarakis et al., Endocrine Reviews, 2006. [10] R. S. Legro et al., Fertility and Sterility, 2013.