Latent State Estimation Method for Students Using Knowledge Tracing Combining with IRT

The Intelligent Tutoring Systems (ITSs) are a system that provides an efficient learning environment by assigning questions that are suitable to the learner's skill state. To improve the performance of ITSs, conventional studies have proposed student models that can estimate skill states with high accuracy. However, these models were based on the assumption that the log data was correct. In other words, they do not take into account the possibility of students cheating or memorizing answers, assuming that the test was conducted fairly and impartially. In recent years, many examinations have been conducted online as a countermeasure against COVID-19 (coronavirus) infections, and students may be able to obtain hints for the examinations using the internet or books. In this way, it is important to know how students approach learning in order to estimate their skill state. In this study, we propose a method for estimating the latent state of learners using a model that combines Knowledge Tracing, the de-facto standard for student modeling method, and Item Response Theory.