Sharapov M. Limitting theorems for estimates of parameters of random processes and fields with long memory.

We estimate the coefficients in regression models with random processes and fields with long memory, investigate their limittig distributions and speed of convergence to these distributions. The methods of spectral analysis are used. Proved reduction theorems make it possible to investigate the models with continuous time instead of those with discrete time and the other way round. The least square estimates and the robust ones are considered; both these estimates are proved to have the equal limitting distributions unlike the case when observations are independent.