(2004). Orthonormality-constrained INDSCAL with nonnegative saliences.
In: Lagan'a, Antonio; Gavrilova, Marina L.; Kumer, Vipin; Mun, Youngsong; Kenneth Tan, C. J. and Gervasi, Osvaldo eds.
Computational Science and Its Applications - ICCSA 2004.
Lecture Notes in Computer Science (3044).
Berlin, Germany: Springer, pp. 952–960.
INDSCAL is a specific model for simultaneous metric multidimensional scaling (MDS) of several data matrices. In the present work the INDSCAL problem is reformulated and studied as a dynamical system on the product manifold of orthonormal and diagonal matrices. The problem for fitting of the INDSCAL model to the data is solved. The resulting algorithms are globally convergent. Numerical examples illustrate their application.
Actions (login may be required)