Hi,
I think your question is old and you already got the answer. I just started dealing with spectral analysis so here is my answer.
first you have some spectral data say m sample in the following order
Sample# wavelegth1 wavelength 2 ....................wavelenght N
1 0.89 0.7 ................................0.1
2 0.7 0.78 .................................0.9
3 . . ................................. .
. . . ................................... .
.
.
m 0.9 0.6 ...................................0.9
remember wavelength 1 means say 450, 0.89, 0.7 etc are corresponding abosorption/reflectance etc.
so you have n*m matrix.
now what you want to look for? is it some kind of material and you want to find some particular element? if yes the just go and od some chamical analysis (wet chemistry) tests and find the exact amount of that element in each sample. store it in a different column.
say we are looking for element "X" content in all the sample, so it will for a column vector (m*1)
you are ready to apply any method say deep learning for classification.
use some of the data as a training set, validation and testing.
hope it help
Best Regards
