“…This is so because being the QSM: Z , by construction nonsingular (otherwise two density functions will be exactly the same), then there always can be computed a QSM inverse: Z −1 , obeying to the usual relationships: Z −1 Z = ZZ −1 = I , in such a way that the trivial result, defining the unknown coefficient vector: will be always obtained within a core set computational scheme. Furthermore, one can retrieve the exact value of the property for any molecule of the core set , just choosing the scalar products: The QSM for several core sets has been used in a quite large set of prediction studies 21–41, in every case employing up to date statistical tools, the usual procedures currently available in classical QSPR studies, see for example references 43–54. The use of the first order fundamental QQSPR equation to construct algorithms, which can be utilized as predictive tools independently of classical QSPR algorithms, has been previously attempted 55, but it has not been continued in practice until recently 42, 56, 57.…”