A method for learning decision tree using genetic algorithm and its application to Kansei engineering system

Abstract

This paper shows the method for learning kansei reasoning rules and its application to canned coffee design. We first obtained the Kansei evaluation data by experiment with canned coffee. Then we analyzed the data by using genetic algorithm. This method extracts Kansei rules to represent the relationship between design elements of canned coffee and Kansei. Extracted rules are compared with results of the conventional statistical method. The results of the method express the design images on the combination design of canned coffee. We show that the acquired rules represent the relationship between human image and the combination of design elements. From the experimental results, the learning method is able to extract the nonlinear relationship among the design elements.