Skip to main content
. 2021 May 4;121(16):9722–9758. doi: 10.1021/acs.chemrev.0c01195

Figure 3.

Figure 3

Illustration on the working principle of the PCA projection. (A) Two-dimensional linear manifold embedded in a high dimensional space. (B) Eigenspectrum of a covariance matrix of the data, as the manifold is two-dimensional a clear gap appears after the second eigenvalues (blue line); a more typical eigenspectrum is shown in orange. (C) Low-dimensional representation of the data obtained through PCA.