Essays on the analysis of market and consumer choice behavior

Abstract

This study proposes new models (i.e., a flexible attribute overlap model and an artificial neural network model) in order to analyze and forecast market and consumer choice behavior. To analyze cigarette market-opening effect and cigarette consumption in Korea, a regression model is applied. In addition, logistic regression model, tobit model, and multinomial logit model are applied to the analysis of cigarette consumers`` characteristics and choice behavior. The limited dependent variable models are very useful to the analysis of situations where dependent variables are either qualitative or limited in their range. In chapter 2, a technique of evaluating substitutability in the portfolio selections of items with multiple attributes is proposed. This approach differs from current techniques in that it evaluates substitutability of portfolios encompassing more than 3 items on the basis of attributes. The values of parameter representing substitutability in the portfolio selection are presented. To examine the validity of the proposed model, a comparative analysis of the predictability of credit card portfolio preferences has been conducted using the proposed method and an existing methodology. The result reveals the superiority of the present approach. The implications of the proposed model for marketing strategies and further research needs are discussed. In chapter 3, a prediction model of consumers choice behavior which utilizes the technique of artificial neural networks is presented. In order to examine the prediction power of the proposed model, a comparative analysis of the predictability of credit card portfolio preferences is conducted using the proposed neural network and the existing multiattribute models of preferences. The results show that the proposed neural net is superior to the multiattribute models in the prediction power. In addition, it has an advantage in the application of the prediction of the choice problem such as the credit card portfo...