Fuzzy homophily in social networks

In the last several years, the study of social networks analysis has gained much interest from the research community. An active research topic is to investigate the structure of social networks, i.e. studying the processes that affect the formation of relationships in a social network. One of the most fundamental notions governing the structure of social networks is homophily. This principle provides us with an illustration of how a network´s surrounding contexts can drive the formation of its links. Given a particular characteristic of interest like religion, gender and individuals´ connections, we concerned to answer to this question "does the network exhibits homophily according to individuals´ characteristic?" For this purpose, we proposed a new method that uses fuzzy concepts to measure social network homophily. In contrast to the traditional approaches, our method handles non-deterministic information in homophily detection process. We applied our approach on a sample dataset drawn from CLOOB social network service. The experimental results demonstrate our approach reaches good efficiency.