A Test of the Equality of SN Ratios Based on the Inverse Gaussian Distribution

Search this article

Abstract

We consider a test of the equality of the signal-to-noise (SN) ratios for measurement systems. We introduce a multiplicative error model and an error term following an inverse Gaussian distribution. Testing the equality of SN ratios is equivalent to testing the equality of coefficients of variation of the inverse Gaussian distribution. We derive new testing procedures via a variance stabilizing transformation and investigate their size and power. We show that these procedures are conservative and have enough power.

Journal

Details 詳細情報について

Report a problem

Back to top