Characterization of atypical virtual campus usage behavior through robust generative relevance analysis

A novel model is introduced that is capable of detecting atypical usage behavior on the cluster structure of the users of a virtual campus, while neutralizing the negative impact of outliers on the clustering process.

Abstract

Virtual campus environments are fastly becoming a mainstream alternative to traditional distance higher education. The Internet medium they use to convey content, also allows the gathering of information on students' online behaviour. The knowledge extracted from this information can be used to fit the educational proposal to the students' needs and requirements. In this study, we introduce a novel model that is capable of detecting atypical usage behavior on the cluster structure of the users of a virtual campus, while neutralizing the negative impact of outliers on the clustering process. This model can simultaneously assess the relative relevance of individual variables on the cluster structure of the users. Experiments carried out on the available data indicate that atypical students' behaviour can be identified and interpreted in terms of those variables which are best at explaining and discriminating their different typologies.