Book Recommendation Service by Improved Association Rule Mining Algorithm

Abstract

With the extensive application of database system, a mass-circulation historical data is accumulated in university library. We applied data mining technology for discovering useful knowledge in circulation data analysis. There are some shortcomings in mining association rules via Apriori algorithm. This paper introduces two methods for improving the efficiency of algorithm, such as filtrating basic item set, or ignoring the transaction records that are useless for frequent items generated. In order to meet the requirement of personal book recommendation service, we applied the improved algorithm to mine association rules from circulation records in university library. A service model is introduced, and may be used for offering recommendation information to the readers. The recommendation model can also be used in other fields, for example, bookstore, information retrieval system, network reference database, etc.