Abstract
It is reported that all items concerning a product, such as brand name, commercial, price, package design and others, which are described in Part 1, give a remarkable effect on forecasting of consumers' purchasing activity. Therefore, in case of forecasting of consumers' purchasing activity, it will be demanded to formulate forecasting problem, considering quantitatively the effects of these all items. This paper presents a probabilistic method to assess the purchasing activity of a new product which is designed by developer with some specified image level of selected items within all items, developing Bayes theory with the likelihood function obtained by the statistical analysis of the selling results of products or the questionnaire surveys. Besides, a probabilistic approach using fuzzy logic as a means to define numerically each image level of all items is formulated and discussed with numerical examples. It is hoped that such a tentative study will help to forecast consumers' purchasing activity for new product with some specified image level of selected items.
