ART1.5SSS for Kansei engineering expert system

Kansei engineering is the technology for translating humans´ feeling into product design. Multivariate analysis is conventional technique for analyze human feeling and building rules. Although these methods are reliable, they are nevertheless incapable of quick analysis and require expertise. These analyses are often used on data that have relatively small sample size for cluster. In this paper, we present ART1.5SSS, a modified version of ART1.5 (Adaptive Resonance Theory) for small sample size clustering. The learning algorithm of ART1.5SSS controls traces to make explanative clusters. The network used for automatic rule building in Kansei engineering expert system, instead of statistical analysis. Categorization performance of new learning rule is compared to multivariate analysis and original ART1.5.