A bayesian based system for evaluating customer satisfaction in an online store

Abstract

This research presents a method of developing Bayesian Network models of complex systems using data sourced from open survey. The process of collection and conversion of random survey information into probability forms are shown with emphasis on integrating expert knowledge and respon-dents’ responses into a Bayesian system model. Random survey questionnaires were administered to customers of an online store to capture their responses about factors that make for efficient product delivery and quality website design. The responses and other known factors that influence online customer satisfaction formed the structures on which the system is modeled, thus formalizing a method of building Bayesian Network models from open survey. Customer satisfaction which is a key performance indicator in businesses is evaluated. Results of simulation showed that Bayesian Networks can be applied in developing models for complex system which can effectively serve as management tool for inference and prediction that can aid decision making.