The previous chapters presented techniques such as artificial neural networks, decision trees, association rules, and clustering used to better understand the present and foresee the possible future. Such techniques, in addition to statistical approaches like regressions, can be described as traditional analytics, often used to solve business problems like customer segmentation, sales and churn prediction, financial forecasting, and fraud and bad-debt detection, among others.
However, other techniques are starting to be utilized that in many ways extend current analytical portfolios. These methods are designed to recognize patterns of behavior and are based on graph analysis, social network analysis, or even complex network analysis. Used to improve business understanding and optimize operational procedures, these newer methods hold great promise to contribute analytic prowess, leading to significant returns for companies in a variety of industries.
This chapter describes graph analysis, its fundamental components, method, objectives, and foundation. Specific, real-life examples of graph analysis application are detailed in the next chapter.
Graph analysis is similar to social network analysis (SNA methods), albeit the former considers all relationships among individuals, permitting flexibility in how both an individual is defined as well as what constitutes a relationship. In this approach an individual could ...