Improvement And Application Of Bayesian Knowledge Tracing Model Based On Cognitive Diagnosis

With the continuous development of the information age,people pay more and more attention to education,and they hope that education can pay attention to every learner.As an important part of learning process,teaching evaluation directly reflects the individual differences of learners.However,the traditional education evaluation can only report general test scores,and fail to provide detailed information about students' knowledge acquisition.Nowadays,people are no longer satisfied with such ambiguous measurement results.They expect to understand the knowledge structure of students in detail on the basis of scores,so as to play an auxiliary teaching function.As a typical student knowledge assessment method,Bayesian knowledge tracing model can accurately reflect a series of personalized data such as learners' learning level and knowledge structure.Based on the existing Bayesian knowledge tracing model,this study adds a parameter matrix of knowledge relationships and proposes CS-BKT model.By using the 2010 KDD data set,the results of the CS-BKT model and the standard Bayesian knowledge tracing model are compared based on four indicators such as accuracy and root mean square error.The experimental results show that the CS-BKT model outperforms the standard knowledge tracing model in all indicators.Based on the CS-BKT model,this study trained and analyzed the sixth-grade mathematics answer data of a primary school,obtained the students' mastery of different knowledge and the level of mathematics ability.Finally,the study generated a personalized cognitive diagnosis report to help teachers better understand students' knowledge level and adjust teaching strategies,so as to improve teaching efficiency.At the same time,the report can help students to grasp their own knowledge structure more clearly and promote the development of personalized learning.