Principle component analysis (PCA)

PCA ref:

• https://www.youtube.com/watch?v=FgakZw6K1QQ (best)
• https://towardsdatascience.com/the-mathematics-behind-principal-component-analysis-fff2d7f4b643 (VVI)
• http://sebastianraschka.com/Articles/2014_pca_step_by_step.html (VVI for pca with python code) 
• https://www.youtube.com/watch?v=xebPVQ1f7nM (vvi pca with python code in video) 
• https://lpsa.swarthmore.edu/MtrxVibe/EigMat/MatrixEigen.html (VVI to undetstand the how to calculate the eigen value and eigen vector) 

benefits:

• reduce the dimension so improve the performence
• try to keep 99% of variance  retained

Application of PCA

• Compression
• reduce the memory/disk need to store data
• speed up the learning alogorithm
•  Visualization
• 2D/3D for visualization 

Before implemening PCA, first try running whatever you want to do with the orginal data/raw data. Only if that does not do what you want, then consiser the PCA.