SPECPROXM

Mapping for computing spectral distances

   W = SPECPROXM(A,TYPE,PAR)
   W = A*SPECPROXM(TYPE,PAR)
   D = B*W

Input
A	Dataset with 1-D spectra
B	Dataset with 1-D spectra
TYPE	Type of spectral distance measure
PAR	Parameter of distance measure

Output
W	Spectral distance mapping
D	Dataset, distance matrix between B and A

Description

This routine is very similar to PROXM, except that it adds a number of  special measures for spectra.

    'L1'     L1 distance, identical to PROXM(A,'m',1);
    'L2'     L2 distance, identical to PROXM(A,'m',2);
    'C1'     L1 distance of the cumulative spectra, after normalization
    'C2'     L2 distance of the cumulative spectra, after normalization
    'D1'     L1 distance of the derivative, after normalization and Gaussian
             blur with st. dev. PAR
    'D2'     L2 distance of the derivative, after normalization and Gaussian
             blur with st. dev. PAR

See also

mappings, datasets, proxm,