DEVELOPMENT OF A LIBRARY BOOK RECOMMENDATION SYSTEM BY ANALYZING USERS’ BEHAVIORS EMPLOYING THE FP-GROWTH ALGORITHM AND USER PROFILE

Abstract

            This research aimed to generate accurate association rules from library borrowing and returning data using the FP-Growth algorithm, to develop a library book recommendation system using the FP-Growth algorithm combined with user profile data, and to study user evaluation with the book recommendation system. The data used to generate association rules were derived from the borrowing and returning history of Ubon Ratchathani University Library during the academic years 2018 – 2022, totaling 6,630 records. These records were processed to generate association rules using the FP-Growth algorithm via the Weka program, with the confidence level of the rules set at 85%. The research consisted of six main steps: (1) data collection, (2) data preparation, (3) generation of association rules using FP-Growth, (4) creation of user profiles for association rule matching, (5) development of the book recommendation system, and (6) evaluation of user satisfaction. The results showed that the association rules generated using FP-Growth could be effectively applied to develop a book recommendation system, helping to suggest books that are likely to match students’ interests. The evaluation of the recommendation system demonstrated a precision of 73.06%, a recall of 85.60%, and an overall F1-score of 78.67%. Furthermore, the user satisfaction assessment of the system usage by forty students, using a questionnaire, showed that the students had an overall satisfaction mean at the highest level ( = 4.77, SD. = 0.32). Based on the satisfaction assessment results, it can be concluded that the developed system is effective and can be practically implemented.