Skip to main content
. 2017 Jan 24;12(1):e0170844. doi: 10.1371/journal.pone.0170844

Fig 5. The results of PLS classification based on spectral dataset in the calibration.

Fig 5

(A) PLS score plot shows that all groups representing different structures of the skin are well-separated along latent factor 1. (B) Prediction result in an additional dataset that are not included for PLS modelling shows that electrical epidermis, normal epidermis and dermis can be distinguishable from each other by using the PLS model. Red and black dot lines show classification threshold for separating the three categories respectively.