Principal Component Analysis in Noise Reduction and Beamforming

Principal component analysis (PCA) is by far the most popular and useful dimensionality reduction technique that one can find in the literature. The objective of PCA is the reduction of the dimension of a random signal vector from M to P, where \(P \ll M\) , with little loss of the useful information. It does so by preserving the variability of the signal as much as possible, where the new variables are uncorrelated. In this chapter, we show how PCA can be applied to noise reduction and beamforming.