A NEW APPROACH TO APPLICATIONS OF THE PATTERN-RECOGNITION METHODS IN ANALYTICAL-CHEMISTRY .1. AUTOMATIC IDENTIFICATION OF STRUCTURAL FRAGMENTS IN ORGANIC-COMPOUNDS

In the paper (being the first part of a coming series) selected applic ations of the results, gathered in a broader cognitive research on an intelligent knowledge engineering program environment, are briefly des cribed. This intelligent programming tool (not discussed here, only th e most important references to it are listed) features an open archite cture, modularity and an idea to use multistrategy learning, multistra tegy knowledge representation and integration of various schemes of kn owledge processing in a single inferential process. These selected app lications, carefully examined on many levels throughout the project, a re exemplified here in the field of the rough-sets pattern recognition , used in the process of automatic learning of spectrum-structure corr elations, followed by discovering of the respective production rules ( by the computer itself), and putting them together into a complete, an alytical knowledgebase for the identification of structural fragments in organic compounds. This new approach of a paramount importance for analytical chemistry, has been developed for six various spectral meth ods (H-1-NMR, C-13-NMR, MS, IR, Raman, and UV). This paper, deals with the infrared applications.