An improvement algorithm for accessing patterns through clustering in interactive VRML environments

Abstract

User's traversal paths in VRML environments often can reveal interesting relationships between the objects. However, the massive objects are always stored and scattered in the storage units. This will increase the search time and reduce the system performance. Unfortunately, this problem is never considered in the traditional VRML environments. In this paper, we develop an efficient clustering method to improve the efficiency of accessing objects. The clustering methodology is particularly appropriate for the exploration of interrelationships among objects to reduce the access time. Based on the co-occurrence table and similarity pattern clustering algorithm, we can cluster these patterns more effectively and efficiently. In order to maintain quality of the clusters, the similarity pattern clustering algorithm is presented which satisfies this require-ment. Our experimental evaluation on the VRML data set shows that our algorithm not only significantly cuts down the access time, but also enhances the computational performance. © Springer-Verlag 2004.