Association based prefetching algorithm in mobile environments

Abstract

In this paper we propose a prefetching algorithm called STAP (Spatial and Temporal Association based Prefetching algorithm). Our methods are based on the analysis of the spatial and temporal associations of the user's request using data mining techniques. First, we exploit an "associative class set" consisting of an itemset of service classes that is close both spatially and temporally and frequently requested together. With the first method, our prefetching algorithm can select a candidate set that is spatially and temporally associated with the previous request of a user. It is shown that through performance experiments STAP is effective in improving system performance. © Springer-Verlag Berlin Heidelberg 2005.