Linear regression with stably distributed residuals

Abstract
A method of obtaining maximum likelihood estimates of the parameters of the stable distribution is developed. The estimation method is both easier to program and more general than estimation methods developed in past research. The approach can handle linear regression with stably distributed residuals. Monte Carlo samples are generated to demonstrate the accuracy of the computer algorithm. The approach works reasonably well for characteristic exponents greater than or equal to 1.70.

This publication has 11 references indexed in Scilit: