An Frame-based Semantic Item Bank Recommendation System Research And Application

With the rapid growth of information technology, cloud education and online tutoring have become a hot spot. As many problems, such as the simple exercise form, online tutoring systems are difficult to do data mining and have low efficiency of learning knowledge, so traditional online education tutoring system is difficult to generalize. Intelligent tutoring system of education can solve those problems. Furthermore, based on human’s cognitive processes, it can analyze user’s potential problems and solve them. Therefore, this thesis solved the problems of complex exercise express, and constructed the recommendation and tutoring system based on analysis of the user’s mastery of the concept and knowledge, mainly about:First, this thesis analyzes the traditional system of education and tutoring in the way students learn the concept of knowledge, and according to the commonalities in cognition, this thesis proposed collaborative filtering algorithm based on domain ontology. This algorithm uses the Analytic Hierarchy Process to compute the Sematic Association Degree between two concepts in a domain ontology. It comprehensively assesses the correlation between concepts, and based on the result, this algorithm can increase the number of the value in user-item evaluation matrix, which can improve the collaborative filtering algorithm’s defect of the matrix data sparsity and cold start.Second, this thesis proposed a method which uses the frame knowledge representation to convert text format exercise to a sematic format. This exercises frame use triples to describe the answer process, which can record the process information of answer efficiently. And the exercise frame, coupled with the domain ontology and process ontology, constitute the semantic exercises bank, which solved the problems of which data structure difficult to represent the complex format and difficult to data mining and reasoning in traditional tutoring system.Last, combined the exercises bank based on frame knowledge representation and Collaborative filtering algorithm based on domain ontology, this thesis designed and implemented a smart learning community system, which is an application of the exercises bank of frame knowledge representation. This system can make students to learn efficiently by dealing with complex format and recovering the derivation of exercises. Based on the data of user interaction, this system analyzes the potential problems of students, and strengthening exercises.