Service demand analysis using multiattribute learning mechanisms

We describe a new approach to analyze customer demand for various types of Internet services and IT systems. We have proposed a multi-attribute learning mechanism called LMDCM (Learning Mechanism using Discrete Choice Models) to evaluate customer satisfaction levels for services. A multiattribute learning mechanism can indicate the customer satisfaction level of each service under given situations. We give an overview of customer-behavior modeling using LMDCM and the framework to analyze customer-churning and service demand. This framework can be used to simulate scenarios under assumed situations. It consists of customer-behavior modeling, service modeling, environment modeling, and scenario simulation functions. Service demand analysis for providers of various services (xSPs) is shown as an application example.