Exploiting semantic annotations for domain-specific entity search

Abstract

Searches on the Web of Data go beyond the retrieval of textual Web sites, and shifts the focus of search engines towards domain-specific entity data, for which the units of retrieval are domain-specific entities instead of textual documents. We study the effect of using semantic annotation in combination with a knowledge graph for domain-specific entity search. Different reasoning, indexing and query-expansion strategies are compared to study their effect in improving the effectiveness of entity search. The results show that the use of semantic annotation and background knowledge can significantly improve the retrieval effectiveness, but require graph structures to be exploited beyond standard reasoning. Our findings can help to develop more effective information and data retrieval methods that can enhance the performance of semantic search engines that operate with structured domain-specific Web data.