A new approach to applications of the pattern recognition methods in analytical chemistry - IV. Automatic identification of structural fragments in organic compounds

In this paper (part of a sequence devoted to automatic identification of organic substructures [1, 2]) a methodology of searching for optional classifiers for selected aromatic fragments embedded in organic molecules is briefly described. The developed methodology uses low-resolution mass spectra and employs, computer program SCANKEE to create the databases of mass spectra and to search them to create spectrum-substructure correlation tables, and finally to convert automatically these tables into the rules data-base which enable effective concluding.