Learning object recommendation model for learners based on learning styles

Abstract

In general, Learning Management Systems (LMS) do not usually “know” the learner and simply present the same learning object to all learners without taking into learner learning styles or their preference. This gave result increased dissatisfied learning object to learners. The main focus of this research is to apply the learning style in learning object recommendation system by using the mapping rules that are developed by word analysis technique. Based on learning style-based design, concept map combination model is proposed to filter unsuitable learning concepts for the course. In part of learning object recommendation, learner model based on Felder and Silverman learning style model is developed to classify learners into 8 styles and implement the compatible value computational methods, which includes three recommendations: i) non-personalized recommendation , ii) preferred feature-based recommendation, and iii) neighbor-based collaborative filtering recommendation. The results show the interesting patterns of the most learners are. It is useful to improve the learning process in the educational system and learning object development. The analysis of preference error (PE) is considered by comparison between actual preferred learning object and compatible prediction, the least error in experimental domain is the feature-based recommendation algorithm.