Application of Techniques of Functional Data Analysis to Spectroscopic Data

A classification and identification problem is solved in the case of functional data. Functional data are data records considered as functions rather than vectors of observations at different values of the argument. This research deals with the spectral lines of rocks, being treated with light in the visible and mid-infrared regions. In a first step Principal Component Analysis in the functional data case will be applied to spectral lines of rock samples of various rock types to find out, whether the spectral lines can be separated.