A learner oriented learning recommendation approach based on mixed concept mapping and immune algorithm

Personalized recommendation in e-learning has attracted the interest of many researchers. How to select the proper learning objects (LOs) and provide a suitable learning path for learners is a complex task. The effectiveness of personalized recommender systems are mostly decided by the reasonable models of learners and learning resources. However, the modeling method needs further research for the learners' special natures in e-learning. Heuristic methods have achieved significant successes on personalized recommendation, but the operators of some heuristic algorithms are often fixed, which diminishes the algorithms' extendibility. In this paper, we propose a learner oriented recommendation approach based on mixed concept mapping and immune algorithm (IA). First, we build universal models for learners and LOs respectively, then apply mixed concept mapping to assimilate their attributes. Second, we model the learner oriented recommendation as a constraint satisfaction problem (CSP) which aims to minimize the penalty function of unsatisfied indexes. Last, we propose an advanced IA which takes the inherent characteristics of personalized recommendation into consideration, and we design the monomer vaccine and block vaccine to optimize the IA. Our approach is compared with other heuristic algorithms and traditional teaching method. From the experimental results, it can be concluded that the proposed approach shows high adaptability and efficiency in e-learning recommendation.