Service choice behavior modeling of telecommunication users

Abstract

With the open use of the communication activities and the diversified applications of services with various functions, a tendency to utilize the communication network in various ways and for various purposes is observed. There is an increasing need to realize network management based on the prediction technique for the communication activities of the user. This paper proposes a modeling for the choice behavior of communication services by applying the disaggregate behavioral model used in the field of urban engineering as a means of estimating commuter demand. The possibility of application to the communication network management is demonstrated. First, the decision‐making factors to be considered in the model are classified and then the general construction procedure for the model is presented. The qualitative and subjective decision‐making factors are represented numerically using the cardinal scaling based on paired comparisons, and are then used as factors of the model. Assuming a communication service providing 28 kinds of information data concerned with daily life activities, a model construction experiment was made. An examination is made based on actual survey data regarding user intention. It is shown that an excellent performance is obtained in the prediction of service choice, learning based on accumulated data, and the evaluation of service functions. Based on the realized performance, the application of the model to the management and operation of the communication network is discussed.