On effective data clustering in bitemporal databases

Abstract

Temporal databases provide built-in supports for efficient recording and querying of time-evolving data. In this paper, data clustering issues in temporal database environment are addressed. Data clustering is one of the most effective techniques that can improve performance of a database system. However, data clustering methods for conventional databases do not perform well in temporal databases because there exist crucial differences between their query patterns. We propose a data clustering measure, called Temporal Affinity, that can be used for the clustering of temporal data. The temporal affinity, which is based on the analysis of query patterns in temporal databases, reflects the closeness of temporal data objects in viewpoints of temporal query processing. We perform experiments to evaluate the proposed measure. The experimental results show that a data clustering method with the temporal affinity works better than other methods.