Paper #539

Títol:

Analysis of matched matrices

Autor:

Michael Greenacre

Data:

Març 2001

Resum:

We consider the joint visualization of two matrices which have common rows and columns, for example multivariate data observed at two time points or split accord-ing to a dichotomous variable. Methods of interest include principal components analysis for interval-scaled data, or correspondence analysis for frequency data or ratio-scaled variables on commensurate scales. A simple result in matrix algebra shows that by setting up the matrices in a particular block format, matrix sum and difference components can be visualized. The case when we have more than two matrices is also discussed and the methodology is applied to data from the International Social Survey Program.

Paraules clau:

Correspondence analysis, International Social Survey Program (ISSP), matched matrices, principal component analysis, singular-value decomposition

Codis JEL:

C19, C88

Àrea de Recerca:

Estadística, Econometria i Mètodes Quantitatius

Publicat a:

Journal of Applied Statistics, 30, 10, (2004), pp. 1101-1113

Descarregar el paper en format PDF