A recommendation system for effective learning strategies: An integrated approach using context-dependent DEA

Universities have been focusing on increasing individualized training and providing appropriate education for students. The individual differences and learning needs of college students should be given enough attention. From the perspective of learning efficiency, we establish a clustering hierarchical progressive improvement model (CHPI), which is based on cluster analysis and context-dependent data envelopment analysis (DEA) methods. The CHPI clusters students' ontological features, employs the context-dependent DEA method to stratify students of different classes, and calculates measures, such as obstacles, to determine the reference path for individuals with inefficient learning processes. The learning strategies are determined according to the gap between the inefficient individual to be improved and the individuals on the reference path. By the study of college English courses as an example, it is found that the CHPI can accurately recommend targeted learning strategies to satisfy the individual needs of college students so that the learning of individuals with inefficient learning processes in a certain stage can be effectively improved. In addition, CHPI can provide specific, efficient suggestions to improve learning efficiency comparing to existing recommendation systems, and has great potential in promoting the integration of education-related researches and expert systems.