Auto-Clustering Algorithm for Heterogeneous Information Network Using Improved Particle Swarm Optimization

• Published in: Applied Mechanics and Materials Vol. 239-240; pp. 1448 - 1455
• Main Authors: Liu, Yang, Liu, Chang Ping, Chen, Jia Shi
• Format: Journal Article
• Language: English
• Published: Zurich Trans Tech Publications Ltd 01.01.2013
• Subjects: Genetic Algorithm (GA); Clustering Algorithm; Knowledge Recovery; Date Mining; Particle Swarm Optimization (PSO)
• ISBN: 9783037855454; 3037855452
• ISSN: 1660-9336; 1662-7482; 1662-7482
• DOI: 10.4028/www.scientific.net/AMM.239-240.1448

NLM (National Library of Medicine) is one heterogeneous information network, which mixes scholars, MeSH (Medical Subject Headings), journals and research domains. Mining the rules and knowledge concealed among NLM is one hot topic in social computing applications. In this paper, an auto-clustering algorithm for NLM was proposed to uncover the embedded knowledge concerned with medical scholars and medical journals. This algorithm adopts particle swarm optimization (PSO) as iterating algorithm to automatically cluster scholars and journals. In addition, our algorithm utilizes the mutation in genetic algorithm (GA) to overcome local optimization, which is one outstanding bottle neck in various heuristic methods. The effectiveness of our algorithm is demonstrated by applying it to a subset of NLM.