A Review of Intelligent Study Recommendation Systems Based on Student Performance

By forecasting student performance and improving study strategies, intelligent study suggestion systems have been developed to improve academic success as educational data becomes more widely available and the need for customized learning experiences increases. These recommendation systems use cutting-edge fields of DM, ML, AI, and data on education analytics to identify variations in student behaviors, academic performance, and engagement levels. After analyzing these parameters, the intelligent recommendation system renders assistance in the mastery of concepts, academic stress, course selection, and many other ways to bridge areas in need of support like lack of personal guidance, ill-fitted from a one-size-fits-all method, or learning gaps in large classrooms. Recent studies from 2021-2025 have evidenced how effective approaches ranging from deep learning model implementations, reinforcement learning frameworks, and hybrid multi-criteria systems to ensemble techniques such as RF give superlative prediction accuracy and adaptability. While these systems hold the promise of personalizing learning pathways and optimizing education-oriented resources, limitations can be observed in terms of data dependence, black-box model interpretability, and scalability with respect to different academic contexts. This review analyzes and compared several approaches studied in the recent literature, highlighting various methods, datasets, and results, stressing the shortcomings of each approach, gaps, and future work in integrating multimodal sources of data, hybrid architectures (symbolic-neural), and adaptive feedback evolving with student progress. The work substantiates intelligent recommendation systems as a basis for the conversion of classical education into a personalized and data-driven learning environment for sustainability.