Process to use multivariate signal responses to analyze a sample

Claim 1

A process for analyzing a sample of unknown constitution by an analytical apparatus utilizing a set of analytical measurements and a set of reference measurements obtained from a set of samples, said process comprising: subjecting said set of samples to measurement using a first analytical apparatus, thereby obtaining a set of analytical measurements comprising a first set of multivariate signal responses; adjusting said first set of multivariate signal responses to remove effects of operational variabilities associated with said first analytical apparatus, thereby obtaining a first matrix of adjusted signal responses; subjecting said set of samples to measurement using a second analytical apparatus, thereby obtaining a set of reference measurements comprising a second set of multivariate signal responses; adjusting said second set of multivariate signal responses to remove effects of operational variabilities associated with said second analytical apparatus, thereby obtaining a second matrix of adjusted signal responses; decomposing one of said matrices of adjusted signal responses into a matrix product that comprises a matrix comprising a set of principal factors and a matrix comprising the amounts associated with each of said set of principal factors; constructing a compressed signal response matrix product that comprises a matrix comprising a subset of said set of principal factors and a matrix comprising the amounts associated with each of the principal factors included in said subset; computing a projection matrix from the matrix comprising the amounts associated with each of the principal factors included in said subset; constructing a first normalized matrix product comprising a normalization matrix and the other of said matrices of adjusted signal responses; constructing a compressed normalized matrix product comprising said projection matrix and said first normalized matrix product; computing values of said normalization matrix that minimize the difference between said first normalized matrix product and said compressed normalized matrix product, thereby producing an optimal normalization matrix; constructing a second normalized matrix product comprising said optimal normalization matrix and said one of said matrices of adjusted signal responses; constructing from said other of said matrices of adjusted signal responses and said second normalized matrix product a calibration useful for analyzing the signal responses obtained from said sample of unknown constitution by said first analytical apparatus; and applying said calibration to analyze the signal responses obtained from said sample of unknown constitution by said first analytical apparatus, thereby enabling the constitution of said sample to be accurately determined.