Determination of the Exact Confidence Intervals for Parameters in a Model of Direct Measurements With Independent Random Errors

Presented is a simple real-data example with analysis of direct pressure measurements by electronic pressure transducer sensors, which illustrates our new method for obtaining exact confidence intervals for the unknown parameters in linear regression models. The method assumes that the errors are expressible as linear combinations of independent random variables with known, however, not necessarily normal distributions. The suggested method for calculating the confidence intervals in linear regression models with possible parameter constraints is based on the application of the characteristic function approach (CFA).