Probably, we think many people have multiple favorite foods. However, there is a bias in food preferences of each person. So, people 's tastes are often diverse.
In this research, we recognize that human preference varies widely as "multimodality" of human preference. And we verified and analyzed about "How much monomorphism (bias) is there in person's preference?" Until now, we conducted verification to analyze the degree of preference for Y shirt design of multiple subjects using various Y shirt designs.

Y-shirt design parts

Y-shirt design example



Result of clustering (Ward method, k-means method, SOM)
Regarding clustering, it is a grouping of data in a nutshell. From these results, we confirmed that some users have the same evaluation tendency as other users.
Related our works
[Conference Presentation] Hiroshi Takenouchi, Masataka Tokumaru, "Verification of Sensibility Representation of Pseudo-Users in Numerical Simulations", Proceedings of the 22nd Workshop on Challenging Ambiguous Feelings, Evaluation Problem Research Group of the Japanese Society for Fuzzy Theory and Intelligent Informatics, pp.78-83, 2017-11 (Kumamoto)(in Japanese).