Recommender system based on data mining: Interlibrary case study

摘要

This paper introduces architecture for recommender system in domain of interlibrary services. The system works based on rules that was extracted from historical usage logs by the aid of data mining techniques. The paper also describes experimental results of pilot implementation of system for Ghadir project (resource sharing program in Iran). The pilot design and implementation phases were data gathering, pre-processing and warehousing, data mining and association rules extraction that all were described in the paper. The logs of 10 years usage of Ghadir project were processed and stored in data warehouse. Then CLOPE algorithm was applied on data to cluster users. After that decision tree technique was used for rule extraction.