Feature Reduction using PCA

I'm working with Gabor filter bank, now I have a huge no.of features by the no.of used filters and I want to deploy the PCA to reduce the features number, but I don't know how to begin and which function to use, plz help.

Answer 1

[Prasanna]

Hi Hend,

It is my understanding that you have extracted a large number of features from your data using a Gabor filter bank, and now you wish to reduce the dimensionality of these features using Principal Component Analysis (PCA).

To perform PCA in MATLAB, you can use the built-in pca function. Suppose your feature matrix is called features, where each row represents a sample and each column represents a feature. You can apply PCA as follows:

[coeff, score, latent] = pca(features);

In the above example, score gives you the transformed features in the new PCA space.

To reduce the number of features, select the first N columns of score that cumulatively explain your desired amount of variance (e.g., 95%). For more information, you can refer the following pca documentation: https://www.mathworks.com/help/stats/pca.html

Hope this helps!