Preprocessing of spectroscopic data to highlight spectral features of materials

Abstract

The study of the extensive data sets generated by spectrometers, which are of the type commonly referred to as big data, plays a crucial role in extracting valuable information on mineral composition in various fields, such as chemistry, geology, archaeology, pharmacy and anthropology. The analysis of these spectroscopic data falls into the category of big data, which requires the application of advanced statistical methods such as principal component analysis and cluster analysis. However, the large amount of data (big data) recorded by spectrometers makes it very difficult to obtain reliable results from raw data. The usual method is to carry out different mathematical transformations of the raw data. Here, we propose to use the affine transformation for highlight the underlying features for each sample. Finally, an application to spectroscopic data collected from minerals or rocks recorded by NASA's Jet Propulsion Laboratory is performed. An illustrative example has been included by analysing three mineral samples, which have different diageneses and parageneses and belong to different mineralogical groups.