A Purely Entity-Based Semantic Search Approach for Document Retrieval

Over the past decade, knowledge bases (KB) have been increasingly utilized to complete and enrich representation of queries documents in order improve document retrieval task. Although many approaches used KB for such purposes, problem how effectively leverage entity-based still needs be resolved. This paper proposes a Purely Entity-based Semantic Search Approach Information Retrieval (PESS4IR) as novel solution. The approach includes (i) its own entity linking method (ii) an inverted indexing method, ranking, (iii) appropriate ranking is designed take advantage all strengths approach. We report findings on performance our approach, which tested by annotated two known tools, REL DBpedia-Spotlight. experiments are performed standard TREC 2004 Robust MSMARCO collections. By using collection, whose terms average annotation scores greater than or equal 0.75, achieves maximum nDCG@5 score (1.00). Also, it shown that PESS4IR alongside another would performance, unless alone those highly queries.