An approximation technique of MLE for the unreplicated linear circular functional relationship model

Abstract

The maximum likelihood estimation (MLE) of parameters for the unreplicated linear circular functional relationship model is discussed in detail. Explanations are given for the difficulty of estimating parameters with no restrictions on the ratio of error concentration parameters. An approximation technique is proposed for the case when the ratio of error concentration parameters is known. The parameter estimates may be obtained iteratively since the closed-form expressions for the maximum likelihood estimates are not available.