Adaptive Lp estimation and its application to photographic astrometry

The P-norm distribution is a general distribution of errors, of which the normal distribution is a particular case with p = 2. In application the best value of p is first estimated from the data and then maximum likelihood estimates are made. The method is applied to a real example of photographic astrometry of Praesepe. The L_p estimates are expected to be better than the LS estimates, when the errors are not normally distributed.