Humans have the ability to perceive an object’s material and properties instantaneously, and use this information to prepare for future actions. Material perception is not only an important factor for humans but also for artificial intelligence robots that are being developed. In addition, material perception is one of the important design requirements in selecting materials suitable for the products desired by consumers and pursued by designers. Because it is impossible to perform material perception using an exact formula, it is determined from tendencies that are identified in surveys. In this study, surveys with a binary selection were conducted, presenting participants with pairs of bipolar adjectives and asking them to choose one of two. After multiple surveys were conducted all the data were merged. Before merging the data, to ensure the reliability of the data homogeneity and correlation were tested using hierarchical clustering, correlation coefficient, and k-means cluster analysis. Afterwards, the merged data was used to analyze universal and comparable perceptual qualities of various material classes using relative frequency and hierarchical cluster analysis.

Abstract
1. 서 론
2. 설문조사 방법
3. 데이터 병합을 위한 데이터 신뢰성 검정
    3.1. 계층적 군집(hierarchical cluster)을 통한 데이터신뢰성 검정
    3.2. 피어슨 상관계수(Pearson correlation coefficient)를 통한 데이터 신뢰성 검정
    3.3. k-평균 군집(k-means cluster)을 통한 데이터 신뢰성 검정
4. 병합된 재료지각 데이터 분석결과
    4.1. 상대빈도를 이용한 재료지각 데이터 분석
    4.2. 계층적 군집을 이용한 재료지각 데이터 분석
5. 고 찰
6. 결 론
References
<저자소개>

• 최재호(국립강릉원주대학교 공과대학 신소재, 생명화학공학부) | Jaeho Choi (Department of Advanced Materials and Biochemical Engineering, Gangneung-Wonju National University, Gangneung 25457, Republic of Korea) Corresponding author