SPECTRAL PARAMETER-ESTIMATION BY AN ITERATIVE QUADRATIC MAXIMUM-LIKELIHOOD METHOD

An iterative quadratic maximum likelihood (IQML) method is applied to spectral parameter estimation of 1D NMR data. A careful comparison of the linear prediction (LP) method based on the singular value decompos ition, the total least squares (TLS) method, and IQML has clearly demo nstrated that IQML is superior to both the LP and TLS methods in terms of the accuracy and bias of the estimation. The superiority of the IQ ML method lies in the fact that constraints on the NMR signal can easi ly be incorporated into the iterative process, The iterative quadratic maximum likelihood method can be used to analyze NMR data directly or to provide a starting point for further data refinement. (C) 1998 Aca demic Press.