การพัฒนาระบบแนะนำข้อมูลสำหรับห้องสมุดออนไลน์โดยใช้วิธีการกรองข้อมูลแบบพึ่งพา ผู้ใช้ร่วมและข้อมูลส่วนบุคคล (THE DEVELOPMENT OF RECOMMENDER SYSTEM FOR E-LIBRARY BY USING COLLABORATIVE FILTERING AND USER PROFILE)

Abstract

The research aimed to design a framework and develop a prototype of the recommender system for E-library by using collaborative filtering and user profile. The dataset were used in this research 1,430 books, level of interest in book groups and 300 users’ preference rates. Development process of this system consisted of two phases: 1) data preparation, matrix table generation, and missing value computation,  2) development of system using PHP language with Microsoft Access database. The system evaluation and development showed that the book can be recommended towards users’ requirement. The system performance was evaluated by precision, recall and F1 measure. The results showed that the precision was 0.83, recall was 0.77 and F1 measure was 0.80. The system was evaluated in term of user satisfaction using questionnaires. The average of user satisfaction towards our system evaluated by 3 specialists was 4.45 with standard deviation of 0.61. The average of user satisfaction towards our system evaluated by 10 general users was 4.50 with standard deviation of 0.60. It can be concluded that the developed system satisfied the users in the good level.