Constructing a User Type Estimation Model

- An Efficient User Type Classification Model Using Decision Tree Analysis -

Abstract

In product design, it has become important to classify users into several types and to design products for each type. In the field of Kansei engineering, users have been classified based on questionnaires. However, questionnaires that have many items require a lot of effort from users. We solved this problem by extracting and modeling key questionnaire items for classification using decision tree analysis. We applied this method to the classification of motorcycle users based on emotion evaluation, and we conducted type estimation. As a result, we developed a method to perform type estimation with a small number of items. The accuracy of the type estimation model was 0.85. In addition, types classified with this method retained the Kansei information of the original rider types, indicating the method’s high validity.