Finding Temporal Influential Users in Social Media Using Association Rule Learning

Abstract

The social media has become an integral part of our daily life. The social web users interact and thus influence each other influence in many aspects. Blogging is one of the most important features of the social web. The bloggers share their views, opinions and ideas in the form of blog posts. The influential bloggers are the leading bloggers who influence the other bloggers in their online communities. The relevant literature presents several studies related to identification of top influential bloggers in last decade. The research domain of finding the top influential bloggers mainly focuses on feature centric models. This research study proposes to apply association rule learning for finding the temporal influential bloggers. The widely used Apriori algorithm is applied using Oracle data miner to find the frequent pattern of bloggers having blog activities together and then we find who influences others based on the rules learned from the association rule mining. The use of standard evaluation measures such as accuracy, precision and F1 score verifies the results. This research study uses the standard dataset of TechCrunch which is a real world blog. The results confirm that the association rule mining can produce rules which help to find the temporal influential bloggers in the blogosphere who are consistent on regular basis. The proposed method achieved accuracy as high as 98% for confidence level of 90%. The identification of the top influential bloggers has enormous applications in advertising, online marketing, e-commerce, promoting a political agenda, influencing elections and affect the government policies.