Design of Case-Based Reasoning System with Community Detection in Complexity Theory

As one of the most hot debated applications of Artificial Intelligence (AI), case-based reasoning (CBR) provides a methodology by reusing stored knowledge into new solutions. Due to its linear storage structure and similarity computation method between the target problem and the stored case, traditional CBR confronts with great challenges in efficiency and flexibility. In this paper, we propose a case-based reasoning system with community detection in complexity theory. With the introduction of Newman algorithm, a cluster tree could be established in order to organize the knowledge base into a hierarchical structure. Based on such design paradigm, the overhead of matching process could be greatly reduced, which will elevate the total performance of the system. Simulations and analysis present the efficiency improvement compared to traditional solutions.