THE DEVELOPMENT OF PREDICTION PROCESS IN SMART LIBRARY SYSTEM BY INTEGRATING ONTOLOGY AND MACHINE LEARNING

Abstract

Real-time analysis, prediction, and recommendation for all library services to members are necessary for smart library development. This process is the important question in this research by having objectives 1) to study the requirement and data structure in smart library system 2) to develop the prediction process in the smart library by integrating ontology and machine learning and 3) to assess the performance of prediction in the smart library by integrating ontology and machine learning. We studied the system requirements from 30 librarians and members and collected data structure in library service transactions by purposive sampling from Academic Resource and Information Technology Center because these universities collected data in electronic database and defined smart library development plan. We developed the new prediction process using the ontology send semantic data to train and predict the classification algorithm. The result of performance assessment founded that K-NN (K=3) can accurately predict the book order at F-score 100%, and the system can provide the predictive oriented recommendation.