Research On The Student Model Based On Neural Fuzzy Comprehensive Method

ICAI is a dominating research in the field of computer aid teaching nowadays, and it represents a kind of novel teaching idea and method: teacher is the simulation object;student is the service object;system based on the state and character of student, tracking the state and character transition and producing teaching information automatic, modulate teaching process and strategy. Otherwise, ICAI is a interdisciplinary subject, involving many branches of computer and education. This paper describes some current researches on the field and proposes an integrated student model, realizes the recognition estimate part of student model using nerve fuzzy comprehensive estimate method.This paper describes the following main researches: analyzing the traditional student model modeling method and finding the lacks of student model;trying to get rid of the lacks of traditional student model which based on the nerve fuzzy comprehensive estimate method. In this paper the student model is separated into five parts: basic information model, safe model, overlay model, recognition model and test estimation model. Using the information model, study history model and recognition estimation model to practice the whole student model function. The recognition estimation model is the main research and is innovate which is based on fuzzy math theory and ANN theory. To produce a novel multistage estimation model by revising traditional fuzzy comprehensive estimation methods.Study shows the proposed student model is more integrated and estimates the students' knowledge state and recognition ability. The estimation method is not rough , it estimate students' state using student recognition situation, character, study environment and study efficiency, the novel estimation method makes estimation be more reasonable. The practice results shows the advantages of the test estimation model based on nerve fuzzy comprehensive estimation method which overcomes the shortcomings of expensive time cost and model subjectivity by fusing the fuzzy comprehensive estimation method and BP network. The nerve fuzzy comprehensiveestimation method is worth to studying.