Is there a way to determine the adequate sample size for discriminant analysis of multivariate data. My data is FTIR spectral data that contains 100 data points for each observation.
