Dimensionality reduction: 1. Feature selection. 2. Feature projection. 2.1 PCA
One of the steps in PCA is to select the top N eigenvectors sorted by eigenvalues. So while it is not feature selection, it can be argued that it involves a feature selection step.