Successive projections algorithm applied to spectral data for the simultaneous determination of flavour enhancers

A novel variable selection strategy for multiple lineal regression (MLR), the successive projections algorithm (SPA), was applied to spectrophotometric data (190–320 nm) for the simultaneous determination of monosodium glutamate (MSG), guanosine-5′-monophosphate (GMP) and inosine-5′-monophosphate (IMP) in dehydrated broths samples. This selection method uses simple operations in a vector space to minimize variable collinearity and has become an interesting variable selection strategy for multivariate calibration. In this work, nine, six and four wavelengths for MSG, GMP and IMP, respectively, were selected to construct calibrations models in order to solve successfully the serious spectral overlapping in samples containing these analytes. The relative errors of prediction (REP) for the validation set were 2.3%, 0.9% and 1.8% for MSG, GMP and IMP, respectively. Commercial samples were analysed and a recovery study was carried out to verify the accuracy of the proposed method with satisfactory results. A continuous flow system was used to develop a simple, cheap and rapid method (sample throughput: 200 h−1), without any previous extraction step.