Question Over PCA:
1. For a Matrix R*C, There are C number of variable and R number of observation. Have skimmed through alot of threads over this medium as well as other.
Alot of places: It is conveyed that the entries in R (basically our observations) Need to be independent. Why that is the need? I do see the literature or PCA derivation specifically state that! What am I missing here?
2. Why do we need to convert matrix to zero mean and unit variance? what would happen if we don't do the same?