<p>Polycystic ovary syndrome (PCOS) is a prevalent endocrine disorder among women of reproductive age, posing considerable health risks. An accurate diagnosis of PCOS relies on recognizing a diverse range of clinical features and applying therapeutic insight, which makes real-time diagnostic evaluation challenging. Consequently, researchers globally have recently explored computer-aided PCOS detection systems as a potential alternative to traditional evaluations. In some studies, Machine Learning algorithms such as Support Vector Machine (SVM), Decision Tree (DT), Logistic Regression (LR), Random Forest (RF), Naïve Bayes (NB), K-Nearest Neighbors (KNN), Gradient Boost (GB), AdaBoost (AB), Extreme Gradient Boosting (XGB), Cat Boosting (CB), and others were used to predict PCOS based on clinical data. In other studies, Deep Learning algorithms, including CNN-based models, U-Net, ResNet, MobileNet, ARU-Net, and others, were employed to predict PCOS using ultrasound images. This Systematic Literature Review (SLR) thoroughly examines a diverse array of techniques for detecting PCOS using both machine learning and deep learning methods. It offers a detailed explanation of the architectures, operational processes, and assesses the proficiency of these models in solving various issues in the medical domains. In the results, the SLR examined thematic connections across various studies, focusing on research profiles, objectives, data characteristics, disease prediction methodologies, and outcomes, while offering a comprehensive summary and critical analysis of machine learning (ML) and deep learning (DL) techniques, with insights into performance metrics, hyperparameter tuning, and model optimization. This systematic review also explores the highly effective and superior ensemble models for PCOS detection. Additionally, it highlights various shortcomings in existing architectures and suggests potential remedies, thereby improving our grasp of their effectiveness and relevance for PCOS assessment processes.</p>