Social Recommender System for Predicting the Needs of Students/Instructors: Review and Proposed Framework

Because of the interest in mining learning from text and even images shared in the posts, comments, and chats of online e-learning networks, it is necessary to identify whether the text contextually reflects the learner´s interests. Thus, based on the posts and comment analysis, it is possible to automatically recommend posts that meet the interests of learners and instructors through an e-learning application system. Until recently, web personalisation in social mining and e-learning mining has received little attention in social networking research. Furthermore, there is little published research in e-learning mining for personalising web content applications. Some progress has been made in addressing the capacity to make recommendations in e-learning and social mining, but current solutions are restricted both by the limitations of the recommendation methods that are based on data mining techniques and by the difficulties with automatic image and text analysis on social networks. This paper describes the merits of a number of common data mining techniques and also reviews the problems associated with their limitation in e-learning and social mining. Such review is necessary since the personalised recommender approaches and techniques represent personalisation services that aim to predict a learner´s interest in some elements existing in the e-learning application systems. From this analysis, it is suggested that such a framework will offer an increased level of instructor and student confidence, support for multi-criteria ratings, and support for semantic data.