HornParallelAnalysi​s(data, K)

To implement Horn (1965) method to determine number of factors after PCA.

Function HornParallelAnalysis.m simulates a distribution of eigenvalues by re-sampling a set of random variables of the real data size from a normal distribution N(0,1), and compares the eigenvalues of the real data and the distribution of eigenvalues from simulation; then the criterion of retaining basis factors is to keep those who are bigger than 95% of simulated distribution of eigenvalues. This is to implement the parallel analysis approach proposed by Horn (1965) and developed by Ledesma et al. (2007). Please read instruction within the function m-file for more information about input and output arguments.