“…Multivariate statistical analysis (MSA) is a popular choice for automated solutions (Bosman et al., 2006; Kannan et al., 2018; Kotula et al., 2003; Malinowski & Howery, 1980; Teng & Gauvin, 2020). Principal component analysis (PCA) and non‐negative matrix factorization (NMF) are two widely used MSA algorithms for the exploration of the HSI‐EDS data (Jany et al., 2017; Kotula et al., 2003; Rossouw et al., 2015, 2016; Teng & Gauvin, 2020). These algorithms aim to extract the underlying features from the available HSI‐EDS data by reducing the dimensionality of the data, where high‐dimensional pixel‐wise data points are linearly projected onto a basis in a low‐dimensional space (Hotelling, 1933; Kotula et al., 2003; Potapov & Lubk, 2019; Tipping & Bishop, 1999).…”