EXPERT SYSTEM BASED ON PRINCIPAL COMPONENT ANALYSIS FOR THE IDENTIFICATION OF MOLECULAR-STRUCTURES FROM VAPOR-PHASE INFRARED-SPECTRA .1. THEORY - IDENTIFICATION OF ALCOHOLS

The concept of an expert system for the determination of molecular structure from a vapor-phase FTIR spectrum using principal component analysis is described. The principal component analysis is applied to a spectral training set of two classes of compounds (namely alcohols and non-alcohols). The scores of the components can be used to describe a discrimination rule that classifies unknowns to either class (i.e., identifies the presence of the alcohol functionality in the structure). Data scaling and weighting techniques applied to the training set spectra are shown to improve the separation power of the first and second principal components. For the example shown, namely testing for the presence of the alcohol functionality, a rule is described that correctly identifies the presence or absence of the alcohol functionality 96% of the time in a validation set. An expert system to derive an estimate of molecular structure solely from vapor-phase infrared spectra could be derived from a series of such rules.