On overview of PCA application strategy in processing high dimensionality forensic data

Principal component analysis (PCA) is very useful for data exploration owing to its unsupervised nature; and has been proven to be a powerful multivariate exploratory tool for processing and interpreting high-dimensional data in many fields such as engineering, physical and biological sciences. In particular, PCA has been applied in various forensic problems to provide either a direct or an indirect solution. Despite the fact that improved variants of the classical PCA are consistently described in the literature, this review concerned only with the classical PCA in view of its overwhelming popularity. This review aims to summarize the application strategy of PCA in solving varying forensic problems. In total, 20 PCA application strategies have been derived from over 80 forensic studies published since 2017. The goal of this paper is to shed light on the versatility and capability of PCA in processing high-dimensional data in general and specifically on its potential in forensic studies. Hence, this work is also relevant to other research communities interested in unravelling latent structure from highdimensional data.

Automated summary

Principal component analysis (PCA) is an unsupervised data exploration tool that has been widely used in engineering, physical, and biological sciences, particularly in forensic issues. Despite the existence of improved variants, this review focuses on the classical PCA. It summarizes the application strategy of PCA in solving forensic problems.