Abstract: The calibration model of a multivariate spectrophotometric system is said to be incorrect, if some interferent impurity component is present in the system and not included in the calibration model.Amethod based on the matrix projection procedure is developed in this paper for testing the calibration model for such system.If the test shows that the analytical system contains an interfering component, the projection vectors of possibly existing in the system are compared with that of the sample mixture to find out the existing interfering species.The experimental data of three analytical samples were analyzed by the proposed method, and satisfactory results were obtained.

Key words: Calibration model testing, Chemometrics, Background correction, Projection algorithm