Gaussian-2 theory as a linear system

We present a procedure for scaling the additive energy terms in the Gaussian-2 (G2) theory by solving a linear system of equations Ax = b using matrix algebra. Our procedure reduces the average absolute deviation (AAD) from 1.20 kcal/mol of the original G2 theory to 0.94 kcal/mol with respect to the G2/91 test set of 125 energies of first-and second-row compounds.

A method for accurately calculating atomization energies, electron affinities, ionization potentials, and proton affinities.