Applying Association Rules for Interesting Recommendations Using Rule Templates

In this paper, we propose a new method of applying association rules for recommendation systems. Association rule algorithms are used to discover associations among items in transaction datasets. However, applying association rule algorithms directly to make recommendations usually generates too many rules; thus, it is difficult to find interesting recommendations for users among so many rules. Rule templates define certain types of rules; therefore, they are one of the interestingness measures that reduce the number of rules that do not interest users. We describe a new method. By defining more appropriate rule templates, we are able to extract interesting rules for users in a recommendation system. Experimental results show that our method increases the accuracy of recommendations.