A fuzzy set model for market share and preference prediction

This paper presents a new method of modelling preferences using fuzzy sets. A fuzzy set conjoint model is developed by incorporating fuzzy measurement of evaluations into the vector preference model. A comparison test of rank prediction is conducted on data gathered from consumer subjects by an adaptive software program. A methodology is developed to generate aggregate crisp market share estimates from individual fuzzy set preference predictions. The result is a substantial improvement in individual predictive validity using the fuzzy set model and more accurate aggregate market share prediction for use in product design, demand forecasting and other management functions.