Improving Library Searches Using Word-Correlation Factors and Folksonomies

Improving Library Searches Using Word-Correlation Factors and Folksonomies Maria Soledad Pera Department of Computer Science Master of Science Libraries, private and public, offer valuable resources to library patrons; however, formulating library queries to retrieve relevant results can be difficult. This occurs because when using a library catalog for library searches, patrons often do not know the exact keywords to be included in a query that match the rigid subject terms (chosen by the Library of Congress) or terms in other fields of a desired library catalog record. These improperly formulated queries often translate into a high percentage of failed searches that retrieve irrelevant results or no results at all. This explains why frustrated library patrons nowadays rely on Web search engines to perform their searches first, and upon obtaining the initial information, such as titles, subject areas, or authors, they query the library catalog. This searching strategy is an evidence of failure of today’s library systems. In solving this problem, we propose an enhanced library system, called EnLibS, which allows partial, similarity matching of (i) tags defined by ordinary users at a folksonomy site which describe the content of books and (ii) keywords in a library query to improve the searches on library catalogs. The proposed library system allows patrons to post a query Q with commonly-used words and ranks the retrieved results according to their degrees of resemblance with Q. Experimental results show that EnLibS (i) reduces the amount of queries that retrieve no results, (ii) obtains high precision in retrieving and accuracy in ranking relevant results, and (iii) achieves a processing time comparable to existing library catalog search engines.