Association rule mining in fuzzy political donor communities

Abstract

Social networks can be found in many domains. While community analysis can help users understand relationships within the network, it can be difficult to analyze the results without considerable effort. However, using the communities to partition underlying data, it is possible to use association rule mining to more easily facilitate meaningful analysis. In this paper we use a real-world dataset drawn from political campaign contributions. The network of donations is treated as a social network and fuzzy hierarchical community detection is applied to the data. The resulting communities are then analyzed with association rule mining to find distinguishing features within the resulting communities. The results show the mined rules help identify notable features for the communities and aid in understanding both shared and differing community characteristics.