Attribute Identification between Spatial Datasets Based on Instance Statistical Similarities

In this paper, we present an approach of using their overlapping instances to derive corresponding attributes between heterogenous spatial datasets for integrating them. The approach consists of three basic steps. Firstly, the overlapping entity instances between heterogeneous spatial datasets are determined by means of geometrical matching. Secondly, attribute similarity, which is expressed by a value between 0 and 1, is computed according to attributes values of the overlapping instance. We design different similarity compute models for different kind of attributes. At last, the corresponding attributes are identified by means of comparing attribute similarity value with the predefined threshold. An experimental result is presented and shows the effectiveness and precision of this approach.