多断面データの図示表現に関する研究(商船・理工論)

概要

論文の詳細を見る
Three-way (or more generally multiway) data analysis has been introduced as the extension of two-way data analysis to higher-order datasets in Tucker [4]. There are several varieties of three-way datasets. Especially, when two of three modes (or ways) in a three-way dataset describe n objects and p-dimensional observational characteristics, and the other mode describes g time samples, this n×p×g dataset X has a form of 3-dimensional array and is called "multi cross-sectional data". In this paper, by using the Tucker's first model (Tucker [5]), we consider 2-dimensional graphical methods for multi cross-sectional data to pursue behaviors of n objects in time. A multi cross-sectional dataset X is approximated in terms of variations of between (shortly, B) and within (shortly, W) groups specified by g time samples. Then, a 2-dimensional B-W plot is proposed and its performance is examined by a real dataset.