Comparison of Interval Analysis and Standard Statistical Methods for Estimating Experimental Data with Uncertainty

Abstract

Interval analysis procedures are used to estimate the parameters of an experimental chemical process under conditions of noise and uncertainty in the probabilistic characteristics of the measurement errors for a small measurement sample. Interval analysis makes it possible to describe exactly the set of admissible values of the estimated parameters needed for correct organization of the technological process through a correct choice of its parameters. Approximate estimates of the parameters are obtained by formal application of a statistical approach and it is shown that in this case the standard statistical approach yields essentially meaningless estimates for the parameters of the process studied here.