Raman spectroscopy with self-organizing feature maps and partial least squares discriminant analysis for discrimination and source correspondence of red gel ink pens

Abstract

• Novel study reporting on discrimination and source correspondence of red gel ink using Self-Organizing Feature Maps (SOFM) and Partial Least Square in tandem with non-destructive techniques of Raman Spectroscopy. • Unsupervised Self-Organising Feature Maps (SOFM) was initially used for data exploration. • Partial Least Square Discriminant Analysis was used as classification model. • Correct prediction rate (100%) for all unknown samples according to specific type of colorants using the trained PLS-DA model.