Spectral classification by generative adversarial linear discriminant analysis

Abstract

Generative adversarial linear discriminant analysis (GALDA) is formulated as a broadly applicable tool for increasing classification accuracy and reducing overfitting in spectrochemical analysis. Although inspired by the successes of generative adversarial neural networks (GANs) for minimizing overfitting artifacts in artificial neural networks, GALDA was built around an independent linear algebra framework distinct from those in GANs. In contrast to feature extraction and data reduction approaches for minimizing overfitting, GALDA performs data augmentation by identifying and adversarially excluding the regions in spectral space in which genuine data do not reside. Relative to non-adversarial analogs, loading plots for dimension reduction showed significant smoothing and more prominent features aligned with spectral peaks following generative adversarial optimization. Classification accuracy was evaluated for GALDA together with other commonly available supervised and unsupervised methods for dimension reduction in simulated spectra generated using an open-source Raman database (Romanian Database of Raman Spectroscopy, RDRS). Spectral analysis was then performed for microscopy measurements of microsphereroids of the blood thinner clopidogrel bisulfate and in THz Raman imaging of common constituents in aspirin tablets. From these collective results, the potential scope of use for GALDA is critically evaluated relative to alternative established spectral dimension reduction and classification methods.