Spring 2016 Brown Bag Speaker Series
Entity Matching in Vector Spatial Data
Dr. Jianhua Wu
Associate Professor and Dean, GIS Department, School of Geography and Environment,
Jiangxi Normal University, China
Visiting Scholar, Spatial Sciences Institute
12:00 – 1:00 p.m.
Friday, February 26, 2016
Allan Hancock Foundation (AHF) B57J
Map conflation is widely used for updating or integrating spatial data from heterogeneous sources. During a conflation process, matching multi-source, multi-scale, and multi-temporal spatial entities is challenging because the corresponding (matched) spatial entities can have significant variations in their geometric shapes, topological structures, positional and geometric uncertainty, attribute details, coding schemes, and semantic expressions. In this presentation I will first introduce the theoretical background related to conflation and entity matching, including classic matching methods and their limitations. Second, I will present our entity matching method that overcomes limitations in existing methods and efficiently handles spatial datasets with large, inconsistent position deviations. I will show our experimental results where the recall, precision, and generality of our method were significantly higher than the traditional methods. Finally, I will discuss future directions of entity matching techniques.
February 26, 2016 @ 12:00 pm - 1:00 pm