The tremendous increase of data sources containing spatial information is bound up with the diversity of geospatial applications such as location-based services (LBS) and global positioning systems. LBS providers use maps to locate spatial entities referring to points of interest (POI), for instance restaurants or locations of events. In our study, we specifically focus on places that tourists can get through LBS. The multiplication of these providers has an impact on the quality of data: spatial entities referring to the same POI may include spatial and terminological properties with incomplete, inconsistent, inaccurate or even wrong data. Thus, entity matching approaches have been proposed to discover correspondences between spatial entities, and experimental validations are traditionally performed to demonstrate the performance of these approaches in terms of effectiveness and efficiency. However, the datasets used in these experiments are rarely available, thus limiting their reuse and a fair comparison between the proposed approaches. In this presentation, We introduce PABench, a benchmark for spatial entity matching, which is available for researchers to assess their algorithms. PABench includes a taxonomy of observed differences, inconsistencies and errors, which enables the characterization of different LBS providers.
Présentation d'un papier accepté à Geoprocessing'2015
