“…These challenges have been addressed by applying various chemometric methods, such as principal component (PCA) 10,[14][15][16][17][18] and linear discriminant analysis (LDA), 14,19,20 least square regression 14,21,22 as well as cluster analysis, 17,23,24 which mostly aim at the extraction of representative spectral features for the characterization and/or classification of the sample. 7,14,17,25,26 Furthermore, traditional SERS analysis, including that in the SERS reporter approach, relies on the comparison of spectra from an unknown sample to either: (i) one or more known reference spectral features from a Raman library/literature or (ii) pure samples of expected components.…”