Dynamic targeting in an online social medium

Laflin, Peter and Mantzaris, Alexander Vassilios and Grindrod, Peter and Ainley, Fiona and Otley, Amanda and Higham, Desmond J. (2012) Dynamic targeting in an online social medium. In: SocInfo 2012, 2012-12-05 - 2012-12-07.

[thumbnail of socinfo2012published.pdf]
Preview
 PDF. Filename: socinfo2012published.pdf
Preprint
Download (1MB)| Preview

Abstract

Online human interactions take place within a dynamic hi- erarchy, where social in uence is determined by qualities such as status, eloquence, trustworthiness, authority and persuasiveness. In this work, we consider topic-based Twitter interaction networks, and address the task of identifying in uential players. Our motivation is the strong desire of many commerical entities to increase their social media presence by engaging positively with pivotal bloggers and tweeters. After discussing some of the issues involved in extracting useful interaction data from a Twitter feed, we dene the concept of an active node subnetwork se- quence. This provides a time-dependent, topic-based, summary of rel- evant Twitter activity. For these types of transient interactions, it has been argued that the ow of information, and hence the in uence of a node, is highly dependent on the timing of the links. Some nodes with relatively small bandwidth may turn out to be key players because of their prescience and their ability to instigate follow-on network activity. To simulate a commercial application, we build an active node subnet- work sequence based on key words in the area of travel and holidays. We then compare a range of network centrality measures, including a recently proposed version that accounts for the arrow of time, with re- spect to their ability to rank important nodes in this dynamic setting. The centrality rankings use only connectivity information (who Tweeted whom, when), but if we post-process the results by examining account details, we nd that the time-respecting, dynamic, approach, which looks at the follow-on ow of information, is less likely to be `misled' by ac- counts that appear to generate large numbers of automatic Tweets with the aim of pushing out web links. We then benchmark these algorith- mically derived rankings against independent feedback from ve social media experts who judge Twitter accounts as part of their professional duties. We nd that the dynamic centrality measures add value to the expert view, and indeed can be hard to distinguish from an expert in terms of who they place in the top ten. We also highlight areas where the algorithmic approach can be rened and improved.

CORE (COnnecting REpositories)