Extracting Learning Features of Knowledge Unit in Knowledge Map

Knowledge unit (KU) is the smallest integral learning object. Extracting learning features of KU (LFKU) is the primary task of intelligent tutoring and personalized e-learning. However, this is a challenging task because LFKUs are a set of intuitive variables. In this paper, we propose a method to automatically extract LFKUs from knowledge map. The method firstly transforms the task into a technical problem of graph calculation based on learning theories of constructivism and knowledge map. Then, based on the theory of complex networks analysis, it regards LFKUs as some state parameters of learning/cognitive process on KU when they walk on knowledge map, so that it extracts LFKUs from topologic information of knowledge map. Finally, our experimental results have shown the soundness of our method.