Methods, systems, apparatus and processes for determining the lithology as well as the mineralogy of subterranean formations surrounding a borehole are described. According to the methods and processes, well log data measurements from neutron spectroscopy applications and associated tool response parameters are solved using an artificial intelligence system, such as an expert system, which in turn generates an appropriate discriminator and/or compositional model that estimates both general and specific lithology as well as the mineralogy constraints of the subterranean formation being analyzed. The methods exhibit good elemental correlation between conventional methods of lithology and mineralogy determination, and can provide numerous output data, including grain density and porosity data within zones of the formation.