Express quality assay of multicomponent mixture

Using simulation modeling we analyzed the method of quality assay of multicomponent mixture (products of chemical and food industries) that applies multisensory systems. We suggested the procedure of sensor rejection with linearly dependent sensibility according to matrix rank of correlation matrix of sensor response. We also proposed the selection of sensors in the system following the eigenvalues of this matrix. The simulated system realizes the discriminatory analysis (supervised pattern recognition) of multicomponent mixtures according to the scheme: (supervision-measuring of sensor responses from the controlled sample-calculation of the probability of their belonging training sample). Data processing is realized in the space of principal components, that increases the probability of correct classification.