Correspondence analysis of an artificial cylinder data (Q1332882)

Based on two circular data sets proposed by \textit{Sh. Iwatsubo} [see Data Analysis and Informatics III, Proc. 3rd Int. Symp., Versailles 1983, 31- 40 (1984; Zbl 0566.62046)] and Otsu (1990), this paper constructs artificial cylinder data to investigate the method of correspondence analysis. The new data consist of the direct products of the two data sets and are of three traits, a circular one and two linear ones. Among these traits, each has its own Guttman effect, giving rise to intensity, closure, involution etc. Furthermore, the phenomenon that one of the traits may be hidden behind the Guttman intensity of the others is discovered. In order to find the substantial traits hidden behind the Guttman effects, a new approach based on eigenvalue analysis is proposed.