消費者の購買地選択行動における選択肢集合の形成方法

Abstract

This paper discusses choice sets in consumer behavior. We have many works analyzing individual choice behavior. These objectives are path choice, household choice, commercial area choice and so forth. Almost all these works take the way that is to express choice behaviors as a mathematical model to analyze and most works assume individuals with the complete information. The assumption of the complete information is the precondition that all individuals have all information about alternatives and relates deeply to individual choice set. In case that there are a large number of decision-makers and alternatives in the area applied the model, the larger size of the area, the larger size of choice set of decision-maker, the assumption, that is decision-maker with complete information, is not valid and gets unreal. There are two ways to consider choice sets in previous papers. The first is the way to determine sets a priori and the next is the probabilistic approach. These approaches are considered multi-steps choice structure. That is the choice an alternative from subset continued from the choice a subset from the set of subsets. The deterministic approach considers that the choice set is consist of available alternatives that the decision makers consider actually. There are a few considerations in terms of the methodology of generation of choice set in the papers that take its approach. The another approach is that the choice probability derives from the joint probability that the choice probability of alternative and subset. There is the representative model that is the Nested Logit model described in Ben-Akiva(1985). The choice sets in the NL model are considered similarity in alternatives in the master choice set therefore it is the subset characterized by similarity. It assumes the individuals with the complete information. In this paper, in the sake of applying the consumer behavior model to a large area such as Tokyo, we at first propose the choice set generation methodology that is deterministic and a priori sets and extend the previous model. The methodology defines the scale of commercial area and distance. The next, we apply the models on Tokyo area and calibrate the models statistically. Then, we assess the proposed model comparing the previous model. Finally, we discuss the distribution of the attractiveness of the choice sets in Tokyo area using the calibrated parameters.