Effectively updating high utility co-location patterns in evolving spatial databases

Abstract

Spatial co-location mining has been used for discovering spatial feature sets which show frequent association relationships based on the spatial neighborhood. In spatial high utility co-location mining, we should consider the utility as a measure of interests, by considering the different value of individual instance that belongs to different feature. This paper presents a problem of updating high utility co-locations on evolving spatial databases which are updated with fresh data at some areas. Updating spatial patterns is a complicated process in that fresh data increase the new neighbor relationships. The increasing of neighbors can affect the result of high utility co-location mining. This paper proposes an algorithm for efficiently updating high utility co-locations and evaluates the algorithm by experiments.