Cross-Language Name Matching for Data Fusion in Linked Open Data

Abstract

--Data quality and accuracy affects the success of data integration in Linked Open Data (LOD). The main goal of data fusionis to represent each real-world entity once on the Web. Data inaccuracy problems exist due to misspelling and a wide range of typographical differences mainly in non-Latin languages, those problems become more complicated when a person is identified by a name, and this name can be presented differently in same/different languages. Up to author's knowledge, the previous approaches which supported Arabic person names are not designed to work with LOD. This paper proposes a framework that uses person names as matching criteria from cross-language LOD Datasets. The proposed framework has substantial improvements in matching results compared to state of the art framework of matching techniques with better matching rate which exceed 6% in precision and 6% in recall.