Impact of Cosmic Rays on the Classification of Raman Spectra via LPA2v and PCA Guilherme Mendes de Andrade, Rogério José Osti, Douglas de Jesus Passoni, Marcos Tadeu Tavares Pacheco, Dorotea Vilanova Garcia, Landulfo Silveira Jr.

Abstract

Raman spectroscopy has been used to analyze several types of substances, however, the existence of noise can interfere with the analysis of Raman peaks in the samples. Using LPA2v and PCA techniques, we sought to assess whether cosmic rays interfere with the interpretation of signals obtained by Raman in gasoline samples without and with the presence of adulterants. The original data was treated, eliminating the fluorescence background. Peaks equivalent to cosmic rays of different intensities and bandwidths were manually added. The two-value annotated paraconsistent logic (LPA2v) and principal component analysis and discriminant analysis (PCA-DA) techniques were applied to data with and without cosmic rays. The classification errors of gasoline samples were 0% in samples without cosmic rays, 6% with strong rays and 13% with large bandwidth rays for the LPA2v technique. For the PCA technique, errors were 6% in samples without cosmic rays, 13% with strong rays and 6% with large bandwidth rays. Both techniques have shown promise for analyzing the influence of cosmic rays on Raman peaks, but both are affected by the existence of these noises.