Business intelligence: fuzzy logic in the risk client analysis

Abstract

The following paper focuses on achieving accurate results through rough data. Using an inference model based on fuzzy logic, human reasoning was proactively stimulated, under certain conditions, in order to deal with the possibility of client-loss due to service quality. The experimentation is carried out by information related of complaint receipts from a period of two years (70,000 registers). For that effect, a prototypical program is set in C++ language, which receives as input the crisp values that result from the failure resolution for each relevant service. The proposed model is intended to classify clients according to the risk they may have in the contractual relationship with the company.