A Data Mining Methodology for Library New Book Recommendation
Customized information service is very important for service provider nowadays. Traditional selective dissemination, as widely discussed in library community requires users¡¦ involvement and only serves a limited amount of users. In this thesis, we propose to employ data mining techniques to discover knowledge in circulation databases so as to provide customized service in library new book recommendation. Our research¡¦s data source is from National Sun Yat-Sen University¡¦s library. We follow a standard data mining procedure and report our experience in this thesis.
Our research uses patron concept hierarchy and book hierarchy with given support threshold and confidence threshold to derived association rules with patron types being antecedent and book types being subsequent. Four algorithms, namely SBSP, SBMP, LatSBMP, MBMP are proposed to facilitate patron and book hierarchy search. Their complexities are compared analytically.
Advisor:Lee-Feng Chien; San-Yih Hwang; Chih-Ping Wei
School:National Sun Yat-Sen University
School Location:China - Taiwan
Source Type:Master's Thesis
Keywords:new book recommendation digital library selective dissemination of information data mining
Date of Publication:07/26/2000