use a method--DTW to match two similar Raman spetros; some metrics and some preprocesses that dealed with origin data.
I sorry to give such a draft here, but I have no time make it clear. However, I still hope I can help you.
The codes belong to some parts of my graduate work. The data I had used in this experiment can't be shared, for they arenot open source.