Linkage of Spatio-Temporal Data and Trajectories

Abstract

The abundance of datasets containing spatio-temporal information calls for novel record linkage methods that can effectively operate on such data to discover records that refer to the same real-world entity. In this paper, we propose the first approach for spatio-temporal record linkage that leverages the power of LSH to provide accuracy guarantees. Through experimental evaluation, we show that our approach outperforms the state-of-the-art method and can scale well to large datasets.