An Object-Oriented Modeling and Implementation of Spatio-Temporal Knowledge Discovery System

Abstract

Spatiotemporal databases deal with changes to spatial objects withtime. The Spatiotemporal knowledge discovery process involves theintegration of such database systems with data warehousing, data miningand online-analytical processing technologies. The applications in thisdomain will process spatial, temporal and attribute data elements anduse spatio-temporal relationships among these elements. Theseapplications deal with the evolution of spatial objects and changes intheir topological relationships, associations along with time. Theseadvanced database applications require storing, management andprocessing of complex spatiotemporal data. In this paper we considerthe classification and modeling requirements for spatiotemporalapplications, System modeling, preprocessing spatiotemporal data,discovering spatiotemporal topological relationships and extension ofapriori algorithm for mining spatiotemporal frequent predicates. Thedesign of an appropriate system that can capture spatiotemporalfeatures of datasets is discussed. Prototype implementation of thesystem is carried out on top of open source object relational spatialdatabase management system called postgresql and postgis. Thealgorithms are experimented on historical cadastral datasets which arecreated using OpenJump. The results that are visualized using OpenJumpsoftware are presented.