Bayesian Student Modelling in a Finnish Language Learning Program

摘要

In this paper we suggest a Bayesian student model for a language learning program, Literate. The Literate program is used to teach young children, just before school age or in the first few years of the elementary school, phoneme-grapheme connections. The Literate program has been developed especially for children who have problems in reading acquisition. The aim of the program is to enhance the phonological awareness of the child playing it. Dysfunctions of phonological awareness have been found to be one of the early signs of dyslexia. We apply our model to empirical data collected using Literate and use Bayesian methodology to compare three different submodels in the light of these data. The inference yielded by the model can be used to make the game adaptive to a particula r student's skills as well as to provide information about the student's learning progress to the teachers and researchers.