Dynamically initializing the student model in a Web-based language tutor

In this paper we describe the method for initializing the student model in a Web-based language tutor. This tutor is an Intelligent Tutoring System (ITS) that operates on the WWW and aims at teaching non-native speakers the domain of the passive voice of the English language. It uses an innovative combination of stereotypes and the distance weighted k-nearest neighbor algorithm to initialize the model of a new student. In particular, the student is first assigned to a stereotype category concerning her/his knowledge level based on her/his performance on a preliminary test. The system then initializes all aspects of the student model using the distance weighted k-nearest neighbor algorithm among the students that belong to the same stereotype category with the new student. The basic idea of the algorithm is to weigh the contribution of each of the neighbor students according to their distance from the new student; the distance between students is calculated based on a similarity measure. In our case the similarity measure is estimated taking into account the students' mother tongue, how careful they are when solving exercises, as well as their knowledge of other languages. This information is acquired directly by the student at her/his first interaction with the system.