COPMOC

Abstract

With the availability of large geo-spatial datasets like maps, repositories of remote-sensing images, location based mobile application data, etc; the concept of spatial data mining is gaining popularity. However, as classical data mining techniques are often inadequate for spatial data mining, different techniques for spatial data mining are being exclusively developed. In spatial data mining, Co-location pattern mining is an important problem which aims at discovering the set of spatial features frequently located together in the geographic proximity. In this paper we propose a framework and algorithm for discovering the frequently occurring co-location patterns of objects in spatial datasets using a co-location mining algorithm which utilizes clustering and map-overlay techniques. The proposed algorithm namely, COPMOC overcomes the limitations of the existing transaction based approaches and the approach which needs distance threshold for neighborhood. The proposed algorithm has been tested and analyzed with a hypothetical dataset generated with 5 different spatial features.