“…To keep a neutral identification (all substances are equiprobable), the sums of weights related to each substance u = 1, ..., R , (∑ i =1,..., T u w u , t ) must finally be forced to be equal. A similar approach, but based on more restrictive assumptions, has been suggested independently for Raman spectra . The authors assume that R is not too large, that the upper bounds of their concentrations are known and that spectral lines are not subjected to any squeeze or shift.…”