Factor Analysis
Factor analysis is a dimentionality reduction technique. It tries to find out the inherent latent factors behind the input features. It is based on the correlation matrix of input features. It requires a large sample size since correlation stabilizes only after large number of data points. It is more exploratory in nature compared to the other dimentionality reduction technique of PCA. It is also different from PCA as PCA is based on explained variance concept of input features. Another difference is PCA only produces orthogonal components. Factor analysis has both variants - oblique (with correlation - direct oblimin, promax) and orthogonal (without correlation - varimax, equimax, quartimax). Factor analysis with oblique rotation is generally preferred as it can also produce orthogonal output if the input features are uncorrelated. Both FA and PCA use variance co-variance matrix for calculation but the diagonal values are different for both. PCA uses 1 in the diagonal values, hen...