Mining Page Category Association on Facebook

Abstract

This paper was proposed to mine association of page category on Facebook. The FP-Growth algorithm co-occurrence is adopted to find frequently of page categories. Then they are used to create association rules. The dataset used to study page category association is the set of page likes of 1,780 users. The minimum support and confidence thresholds are investigated to fine the appropriate values for generating the rules. From the investigation, the appropriate minimum sup-port and confidence values are 10% and 50%, respectively. They give 93.71% of accuracy and use only 892 rules from 7,749 rules for prediction. The association rules can be used to plan or propose pages to interested users.