Cluster-based FAQ retrieval using latent term weights

초록

Web service researchers have introduced a cluster-based retrieval system, the FAQ Retrieval and Clustering Technique (FRACT), that generates clusters of similar query logs. FRACT acquires users' query logs and classifies their contents into FAQ categories. It uses the query-log clusters, which include many terms for managing lexical chasms between the query and the FAQs. It uses high-level knowledge bases that are well-organized in specific domains. It applies latent semantic analysis (LSA), which is a statistical computation method of extracting and representing the contextual-usage meaning of words. It consists of two subsystems, including query-log clustering and cluster-based retrieval. The query-log clustering presents FAQs and query logs in latent semantic space. The cluster-based retrieval utilizes the query-log clusters as a document smoothing during information retrieval.

키워드