Association rule based clustering of electronic resources in university digital library

Abstract

Library Analytics is used to analyze the huge amount of data that is collected by most colleges and universities when the library electronic resources are browsed. In this research work, we have analyzed the library usage data to accomplish the task of e-resource item clustering. We have compared different clustering algorithms and found that association-rule (ARM) based clustering is more accurate than others and it also identifies the hidden relationships between articles which are content-wise not similar. We have also shown that items in the same cluster offer a good source for recommendation.