Abstract: The term Technology Enhanced Learning (TEL) is gradually familiar to both researchers in E-Learning and learners. This development aims to facilitate learners in searching for suitable learning resources such as courses, learning contents which can satisfy those learner’s needs or interests. In general, the current techniques of recommender system (RS) play a major role in developing such education applications. There are currently two remarkable trends in building a recommendation system, including collaborative based RS and content based RS. Particularly, each approach employs some different algorithms for implementation depending on applied domains. In this paper, the logistic regression classification is analyzed to design a collaborative filtering (CF) recommendation system for courses in formal training programs where students could be advised to choose some suitable courses to their preferences in an upcoming semester, basing on ratings from previous students who finish the same training program. In addition, the problem of missing values is discussed in detailed. Generally, the purpose of this study is to propose a suitable method for building a CF recommendation system in course domain.

Keywords: Course recommendation systems, Collaborative Filtering, Gaussian Distribution, Missing value, Mean imputation, multivariate Gaussian distribution. Logistic regression.