Cluster Analysis for Users' Modeling in Intelligent E-Learning Systems

In the paper, the data driven approach for users' modeling in intelligent e-learning system is considered. Individual models are based on preferred learning styles dimensions, according to which students focus on different types of information and show different performances in educational process. Building individual models of learners allows for adjusting teaching paths and materials into their needs. In the presented approach, students are divided into groups by unsupervised classification. Application of two-phase hierarchical clustering algorithm which enables tutors to determine such parameters as maximal number of groups, clustering threshold and weights for different learning style dimensions is described. Experimental results connected with modeling real groups of students are discussed.