Choosing between GC-FTIR and GC-MS spectra for an efficient intelligent identification of illicit amphetamines

摘要

In this paper we are presenting a comparative analysis between several expert systems built for the identification of illicit amphetamines based on their GC-MR and GC-MS spectra. The systems were built using Artificial Neural Networks (ANNs), and are dedicated to the recognition of amphetamines. Structure-activity relationships are incorporated into the knowledge base, allowing the systems to identify the amphetamines according to their toxicological activity (stimulant or hallucinogenic). The results show that GC-FTIR data are much more relevant for the efficiency of the expert systems, probably due to the fact that these spectra constitute a "fingerprint" of the molecular structures. We are also presenting a spectroscopic analysis in order to evaluate the relevance of each type of input variable (absorption and abundance) on which the recognition of an unknown sample is based.