On the Estimation of TSP Distribution Parameters Based on Experimental Data

For the family of two-sided power distributions, various methods of parameter estimation based on experimental data are considered: the maximum likelihood method, the method of moments, the method of choosing the most fitting distribution from the family (based on the inverse mapping). Biases of parameter estimates obtained using these methods are estimated; correction factors are obtained for maximum likelihood estimates. The efficiency of various estimates is also compared. Some other typical problems are considered: estimation of the standard deviation of a random variable belonging to the TSP family (including obtaining correction factors to obtain unbiased estimates), estimation of the position parameter. Restrictions are imposed on the possible values of the distribution parameters to be determined based on the available sample data.