×

Some Euclidean space properties applied to data analysis. (English) Zbl 0592.62046

Summary: The principal components analysis (PCA) gives good results in several fields, but the factorizing is unique (function of the correlation coefficient matrix). By using orthonormal bases in Euclidean spaces we generalize the PCA, making the interpretation of results easier (by geometric considerations). It is possible as well to factorize the data according to the operator’s aims, keeping the weight calculation (for any factor and for individual projections on the factor).

MSC:

62H25 Factor analysis and principal components; correspondence analysis
62-07 Data analysis (statistics) (MSC2010)
PDFBibTeX XMLCite
Full Text: EuDML