A Systematic Review of Ai For Polycystic Ovary Syndrome Deduction and Prediction

Abstract

Polycystic ovary syndrome(PCOS) is a complex endocrine disorder that affects approximately 5-10% of women of reproductive age, giving to infertility, metabolic diseases, and psychological distress. The diverse symptoms and varying disease progression scales make accurate diagnosis difficult. This study proposes a cognitive artificial intelligence (AI)–based approach for the deduction and prediction of PCOS by integrating clinical, biochemical, and lifestyle data into an intelligent decision-support framework. This systematic review evaluates recent AI approaches—machine learning (ML), deep learning (DL), explainable AI (XAI), and large language models—for improving diagnosis, predicting risk, discovering biomarkers, and supporting clinical decisions. Explainable AI techniques are incorporated to enhance transparency and clinical interpretability of model decisions. Experimental evaluation demonstrates improved prediction accuracy and robustness compared to traditional diagnostic methods, highlighting the potential of cognitive AI to support clinicians in early detection and personalized management of PCOS. This approach aims to reduce diagnostic delays, improve patient outcomes, and contribute to intelligent healthcare systems for women’s health. The accuracy of a cognitive AI approach for PCOS deduction and prediction is based on the different dataset. However, based on published AI/ML-based PCOS studies and typical cognitive-AI systems, accuracy of ranges from 85% to92% . Traditional PCOS diagnostic approaches face significant limitations due to symptom variability, incomplete patient data, and overlapping clinical indicators with other endocrine disorders. These challenges often lead to delayed or inaccurate diagnosis. To overcome this problem, cognitive reasoning techniques are employed in the proposed AI-based PCOS diagnosis and prediction system.