An Efficient Clustering Algorithm for Patterns Placement in Walkthrough System

Abstract

Mining of sequential patterns in walkthrough systems is an interesting data mining problem. It can be envisioned as a tool for forecasting and prediction of the future behavior of user’s traversal patterns. In the past, how to display the object faster in the next time were their concerns. They seldom consider the problem of access times of objects in the storage systems. In this paper, we will consider this problem and solve this by clustering. Clustering methodology is particularly appropriate for the exploration of interrelationships among objects to reduce the access times of objects. We record the user’s path as log-data and store it in database. After a certain period of time, we will process the log-data database for user traversal paths and find out their characteristics, which will be utilized to determine the optimal physical organization of those VRML objects on disks. Meanwhile, we also introduce the relationships among transactions, views and objects. According to these relationships, we suggest the clustering criteria--inter-pattern similarity, which utilize these characteristics to distribute the objects into the appropriate clusters. As a result, the large-scale VRML models could be accessed more efficiently, allowing for a real-time walk-through in the scene.