The algorithm of substructure recognition in organic compounds by the artificial intelligence method on the basis of IR-spectra

An algorithm for identification of the IR-active substructures has been developed. In the algorithm, which recognizes about 100 substructures, the analysis is based on three parameters for a set of characteristic bands: position, intensity and arbitrarily chosen diagnostics. The algorithm provides two ways of identification: a. standard method, which compares an IR spectrum with the correlation tables contained in the algorithm, b. dynamic method advanced here which consists in that the recognized substructures cause blocking of their characteristic bands making them almost inaccessible for further analysis.