A soft-computing-based method for the automatic discovery of fuzzy rules in databases: Uses for academic research and management support in marketing

a r t i c l e i n f o The study here highlights the potential that analytical methods based on Knowledge Discovery in Databases (KDD) methodologies have to aid both the resolution of unstructured marketing/business problems and the process of scholarly knowledge discovery. The authors present and discuss the application of KDD in these situations prior to the presentation of an analytical method based on fuzzy logic and evolutionary algorithms, developed to analyze marketing databases and uncover relationships among variables. A detailed implementation on a pre-existing data set illustrates the method. Since the 1980s, increasing complexity, uncertainty, unpredict-ability, and ill-definition characterize competitive environments of companies worldwide (Hough & Duffy, 1987; Hunt, 2000). To be competitive, firms need now to develop adequate competences to effectively compete in a turbulent context (Grant, 2003; Lei, Hitt, & Bettis, 1996). These critical competencies include having the capability to process huge amounts of data, and generating and disseminating pertinent knowledge to aid managers' decision making (Sher & Lee, 2004). Notwithstanding, most firms' research and analytical methods are unable to cope with the informational needs of these new competitive environments (Tay & Lusch, 2005). Organizations now demand research and analytical methods that provide added-value information to support strategic and operational decisions (Avison, Eardley, & Powel, 1998). A widely accepted term, integrating the diversity of tools and computerized systems used by firms for this purpose, is marketing The evolution of MkMSS parallels an evolution in the type of decision problems managers must address. Specifically, while some marketing problems are well-structured, many others, especially those belonging to the strategic sphere, are complex, ill-defined, or unstructured (Wierenga, 2010). Such unstructured problems contain elements, and/or relationships between elements, that are not known a priori. These problems are novel by definition, and therefore managers cannot depend on a pre-programmed decision protocol to solve such problems (Wierenga & Van Bruggen, 1997, 2000). For instance , a sales manager could receive a large number of customer complaints reporting unethical salesperson behavior. The novelty of a problem of this kind needs the manager to take an open approach to understanding the possible explanatory factors, and to subsequently develop a unique solution. Ill-defined problems such as this frequently require systems able to both work with and provide outputs expressed in qualitative terms, because managers must often use qualitative judgmental processes in these cases. Such systems form a specific type of MkMSS category (i.e., …