A Trial to Support to Understand Decomposition Methods for Multidimensional Data by Solving 3-D Puzzles
山本, 直樹 ,
石田, 明男 ,
平田, 将大村上, 純
熊本高等専門学校 研究紀要 = RESEARCH REPORTS OF KUMAMOTO NATIONAL COLLEGE OF TECHNOLOGY
74 , 2018-01 , 独立行政法人国立高等専門学校機構 熊本高等専門学校
We have studied about multidimensional data analysis, and the related themes with this have been given to our students in their graduation research. Although the higher-order singular value decomposition (HOSVD) method is often used to treat multidimensional data, such as many big data processing problem, it is difficult for the students to understand the fundamental principle of it. In order to improve this situation, as our previous work, we developed an understanding support system of HOSVD by visualizing its calculation process. This time, we tried to give three-dimensional (3-D) puzzles to the students to solve by using HOSVD and to let them study that principle with interest. We describe our trial in this paper.