A Hybrid Approach for Knowledge Recommendation

Abstract

Knowledge sharing is critical to knowledge management as it enables employees to share their knowledge. However, knowledge searching is a very time-consuming work. Additionally, in the context of an unsolved puzzle or unknown task, users typically have to determine the knowledge for which they will search. Therefore, knowledge management platforms for enterprises should have knowledge recommendation functionality. Hybrid recommendation systems (RS) have been developed to overcome, or at least to mitigate, the limitations of collaborative filtering. Because Genetic Algorithm (GA) is good at searching, it can cluster data according to similarities. However, the increase in the amount of data and information reduces the performance of a GA, thereby increasing cost of finding a solution. This work applies a novel method for incorporating a GA and rough set theory into clustering. In this paper, this work presents a hybrid knowledge recommendation model, which has a two-phase model for clustering and recommending. Approach implementation is demonstrated, as are its effectiveness and efficiency.