A RECOMMENDER SYSTEM BASED ON ONE-CLASS CLASSIFICATION

Abstract

In the early days, outliers were considered anomalous,  possibly  erroneous observations, which should be identified and removed from the analysis, when building the model. However, analyzing such outliers might be useful in some case.

In this paper, we develop a recommender system to identify and explain ab-normal accesses to our system, the Online Learning System – based on Moodle (OLS) and the online Educational Management System (EMS). Our approach is proposed on the basic of one-class classification and Support Vector Machine to process and classify raw text data. Our experiments demonstrate the utility and accuracy of the system in dataset from the usage of our system.