Parallel analysis (Houts, 1965) diagnoses the number of dimension to approximate data.
parallel simulate eigen values of the random matrix.
fit = parallel(data, 1000, f = fa); # Return `Parallel` struct
Type of data is Matrix or DataFrame and its elements, by the default argument, are assumed to be ordered categorical. If data has continuous variables, ...f = x -> fa(x, cor_method = :Pearson) has to be used.
fit = parallel(data, 1000, f = x -> fa(x, cor_method = :Pearson));
Plot recipe for Parallel is implemented.
using Plots
plot(fit)