TESTING THE HYPOTHESIS ON THE DISTRIBUTION OF THE STATE-VECTOR OF A SYSTEM IN REGRESSION EXPERIMENTS UNDER MULTICOLLINEARITY CONDITIONS

A method is designed for testing the hypothesis concerning the magnitu des of variance and mean of a linear function of the state vector unde r two possible values for the useful parameter and constraints on inte rfering parameters. The observation model is defined as regression on the random state vector. The results recorded in a series of repeated experiments form a sample of the observed vector. The state vector and noise vector are assumed to be normally distributed.