Distribution transformation

nvenkovic - Posted on 01 September 2011

Hello, I have several questions about transformation towards and back-transformation from standard normal deviates obtained in the frame of Gaussian simulations. 1) I assume the hard data I use from my experimental design follow a Beta distribution. Is there any straightforward process to transform those data with respect to a standard normal distribution? 2) My simulated data are not exactly standard normal deviates. Should I consider the exact simulated distributions rather than standard normal pdf's to back-transform? 3) If so, should I consider 1 simulated pdf per location (grid point) or rather a sole distribution for all the spatial random variable? 4) Is there any straightforward process to back-transform a standard normal deviate - or its simulated approximation - into a given beta distribution? Best regards, Nicolas.