Instance Coreference Resolution in Multi-ontology Linked Data Resources

摘要: Web of linked data is one the main principles for realization semantic web ideals. In recent years, different providers have produced many sources in Linking Open Data (LOD) cloud upon schemas. Isolated published are not themselves so beneficial intelligent applications and agents context web. It possible to take advantage potential capacity without integrating various sources. The challenge integration limited instances; rather, schema heterogeneity affects discovering instances with same identity. this paper we propose a novel approach, SBUEI, instance co-reference resolution between even heterogeneous For purpose, SBUEI considers entity problem both levels. process matching applied levels consecutively let system discover identical instances. also applies new approach consolidation level. After finding instances, searches locally around them order find more that equal. Experiments show obtains promising results high precision recall.