| Summary: | 博士 === 國立交通大學 === 科技管理研究所 === 97 === Under the dynamic environment, consumers have much more chance to access products and media; this may create more difference in their preferences. This study aims to improve the elder method of market segmentation; first, lifestyle variables are used to predict consumer choices to increase the richness of the explanation. Second, the novel model which integrates general lifestyle variables and appropriate methodologies is proposed to solve the complexity of the previous research problems. To simplify and reduce the length of the questionnaire in lifestyle research, classification and regression tree (CART) is applied to discover the general lifestyle variables to segment the market. The empirical results show that, under different purchasing situation, the explanatory power of general lifestyle variables is not less than those traditional demographic variables; in fact, some models with different combinations of general lifestyle variables have higher explanatory power than demographic ones. Also, in this paper, a novel method, integrated by CART and rough sets, is propose to improve the shortcoming of CART. From the comparison of CART, rough sets and the proposed method, we can conclude that the proposed method appropriately integrates the advantages of CART and rough sets.
|
|---|
