Learning Concept Recommendation based on Sequential Pattern Mining

Sequential pattern mining is new trend in data mining domain with many useful applications, especially commercial application but it also results surprised effect in adaptive learning. Suppose there is an adaptive e-learning website, a student access learning material / do exercises relating domain concepts in sessions. His learning sequences which are lists of concepts accessed after total study sessions construct the learning sequence database S. S is mined to find the sequences which are expected to be learned frequently or preferred by student. Such sequences called sequential patterns are use to recommend appropriate concepts / learning objects to students in his next visits. It results in enhancing the quality of adaptive learning system. This process is sequential pattern mining. In paper, we also suppose an approach to break sequential pattern s=square c1, c2,..., cm square into association rules including left-hand and right-hand in form ci -> cj. Left-hand is considered as source concept, right-hand is treated as recommended concept available to students.