ABSTRACT
The basic hypothesis of chemistry, which extends to physics, biology, medicine, and other natural and even social sciences, is that similar structures (or similar systems) have similar properties. Thus, searching for regularities in structure-property data becomes one of the central topics of the natural sciences and quantication of the structure-property data one of the major challenges. Consider, for example, a set of experimental data, such as the boiling points of alkanes [1,2], the chromatographic retention data [3,4], the C13 chemical shifts in alkanes [5], the diamagnetic susceptibilities of alkanes and alkenes [6], the determination of pKa values of selected organic acids [7], the reaction rate constants of volatile unsaturated hydrocarbons with OH radicals [8], the resonance energies and aromaticities of conjugated hydrocarbons [9-33], the ring currents in non-benzenoid hydrocarbons [34-38], the activity of nonspecic anesthesia [39], the mutagenicity of nitroarenes [40], the taste of sulfamates [41], the similarity in benzomorphans [42], the quantitative structure-activity relationship of avonoid p56lck protein tyrosine kinase inhibitors by neural network approach [43], or the characterization of trans-membrane proteins [44]. There is no end to the list; we just mention a selection of topics with which the present authors are more familiar.
