Facebook, Linkedin, Myspace, and other social networks have become a very important environment in which people interact, exchange information about products, services, movies and music, and so forth. New trends and hot items rapidly move through these networks. Clearly, modern marketing has to focus on the possibilities of taking advantage of these networks. The determination of people who are leaders and trendsetters within a social network would be a great benefit for marketing. In recent papers, the authors have developed a model of social networks based on the use of fuzzy set theory and other soft granular computing technologies. This is called the Framework for Intelligent Social Network Analysis (FISNA). Using granular computing, the authors express concepts associated with social networks in a human-focused manner. Since human beings predominantly use linguistic terms in order to communicate, reason, and understand, they are able to build bridges between human conceptualization and the formal mathematical representation of the social networks. Consider, for example, a concept such as “leader.” An analyst may be able to express, in linguistic terms, using a network relevant vocabulary, the properties of a leader. The authors’ framework enables translation of this linguistic description into a mathematical formalism that allows for determination of how true it is that a particular person, a node in the network, is a leader. The authors use fuzzy set methodologies, and more generally granular computing, to provide the necessary bridge between the human analyst and the formal model of the network. In this chapter, the authors investigate and describe the use of the FISNA technology to help in the modeling of market related concepts in social networks.
TopIntroduction
A social network provides a vast amount of information in the form of a social graph of human relationships, interactions and behaviors. The usefulness of this information to advertisers and marketers is readily apparent. For example, identifying people who are high influencer of their social circle, known as opinion leaders and trendsetters, can be useful in the preliminary steps of preparing a marketing plan and strategy (Doyle, 2007). In the pharmaceutical industry, key opinion leaders are medical professionals whose opinions are frequently consulted for decisions in products and treatments (Nair, Manchanda & Bhatia, 2010). Marketers are also interested in setting apart followers and trendsetters for music, fashion, and media; the buying patterns of trendsetters provide insights for future trends, directly and indirectly promotes the sale of products (Maldonado, 2010). The complex inter-relational structure of these networks greatly complicates the task of extracting the kinds of information desired by marketers. Our goal here is to provide a language which can be used to intelligently query a social network. Here we shall describe fuzzy set operators developed by Yager (2008, 2010a, 2010b) which can be used to develop human focused network information retrieval techniques.
Considerable recent interest has been focused on social relational network analysis (Carrington, Scott & Wasserman, 2007). A notable example of this has been in the analysis of terrorist and criminal organizations (Popp & Yen, 2006). Another area of applicability is in the domain of social networks such as Facebook, Myspace and LinkedIn that are rapidly gaining importance on the Internet.
In trying to extend our capabilities to analyze social relational networks, an important objective is to associate human concepts and ideas with these networks. Since human beings predominantly use linguistic terms, in which to communicate, reason and understand, we become faced with the task of trying to build bridges between human conceptualization and the formal mathematical representation of the social network.
Consider for example a concept, such as “leader”. An analyst may be able to express, in linguistic terms, using a network relevant vocabulary, properties of a leader. Our task then becomes a translation of this linguistic description into a mathematical formalism that allows us to determine how true it is that a particular node is a leader.
In this work we began looking at the possibility of using fuzzy set methodologies and more generally granular computing (Zadeh, 1998; Lin, Yao & Zadeh 2002; Bargiela & Pedrycz 2003; Yager, 2006) to provide the necessary bridge between the human analyst and the formal model of the network.
Our interest in focusing on this technology is based on the confluence of two important factors. One of these is that fuzzy set theory and particularly Zadeh's paradigm of computer with words (Zadeh, 1996, 1999) which was especially developed for the task of representing human linguistic concepts in terms of a mathematical object, a fuzzy subset. Fuzzy logic has large repertoire of operations that allows for the combination of these sets in ways that mimic the logic of human reasoning and deduction. The second important factor is the nature of the formal mathematical model of social networks. The standard formal model used to represent a social network is a mathematical structure called a relationship. Using this structure, the members of the network constitute a set of elements, the connections in a network are represented as pairs of elements, and the network is viewed as the set of all these pairs. The key observation here is that the standard form of network representatives is in terms of set theory. The fact that the underlying representation of the social network is in set theoretic terms makes it well-suited to a marriage with the fuzzy set approach. In Figure 1 we show the FISNA, Framework for Intelligent Social Network Analysis.
Figure 1. Framework for Intelligent Social Network Analysis (FISNA)