Chinese Information Retrieval Based on Terms and Ontology
Lingpeng Yang, Donghong Ji, Li Tang · NTCIR · 2004
In this paper, we describe our approach for single language information retrieval (SLIR) on Chinese language of NTCIR4 tasks. Firstly, we automatically extract terms (short-terms and long terms) from document set and use them to build indexes; secondly, for a query, we use short terms in the query and documents to do initial retrieval; thirdly, we build an ontology for the query to do query expansion and implement second retrieval. Finally, we use long terms to reorder the top N retrieved documents. Experiments show that the method achieves good results for both T-run and D-Run SLIR tasks of Chinese language.