Suggesting Subject Headings Using Web Information Sources

We proposed a method that suggests subject headings based on user queries when a pattern-matching algorithm fails to locate subject searches for Online Public Access Catalogs (OPAC). We combined information obtained from Wikipedia, Amazon, and Google for query expansion. Our method has two main advantages: (1) availability for any library without customizing OPACs, and (2) ability to suggest subject headings when a query string is not included in OPAC’s bibliographic information. Three experimental results using computer terms revealed the following: (1) Suggested subject headings were related to the input term; (2) Suggested subject headings were better when we used a mixture of Wikipedia, Amazon, and Google than when just using one of them; (3) Our method can suggest subject headings when OPAC mining HIROSHI UEDA, HARUMI MURAKAMI and SHOJI TATSUMI 2 cannot. We conclude that our method can serve as an alternative when pattern-matching algorithms fail.