An Efficient Mining and Clustering Algorithm for Interactive Walk-Through Traversal Patterns

Many algorithms are suggested for improving the performance of a walkthrough system which contains large-scale VRML models. Since massive objects are stored in the storage systems, and may be scattered, this situation increases the search time to access the objects. However, traditional walkthrough system never considers the problem of how to reduce access times of objects in the storage systems. The quality of the walkthrough system needs to be improved in order to meet the user´s demand. In this paper, we present an efficient mining method to improve the efficiency of object accesses. Meanwhile, clustering methodology is particularly appropriate for the exploration of interrelationships among objects to reduce the access time. In other words, prediction and accuracy are our major concerns for improving the system performance. Also, we introduce the relationship measures among transactions, views and objects. Based on these relationship measures, the clustering algorithm will determine how to cluster and the optimal physical organization of those VRML objects on disks. Besides, we suggest two clustering criteria - intra-pattern similarity matrix and inter-pattern frequency table. Our experimental evaluation on the walkthrough data set shows that our algorithm doesn´t only significantly cut down the access time, but also enhance the accuracy of data prefetch.