You are here:
APPLICATIONS OF INFORMATION THEORY AND PATTERN RECOGNITION TO GAS CHROMATOGRAPHY-MASS SPECTROMETRIC ANALYSIS OF TOXIC ORGANIC COMPOUNDS IN AMBIENT AIR
Citation:
Scott, D. APPLICATIONS OF INFORMATION THEORY AND PATTERN RECOGNITION TO GAS CHROMATOGRAPHY-MASS SPECTROMETRIC ANALYSIS OF TOXIC ORGANIC COMPOUNDS IN AMBIENT AIR. U.S. Environmental Protection Agency, Washington, D.C., EPA/600/D-86/103 (NTIS PB86196706).
Description:
The number of information channels contained in the gas chromatographic, mass spectrometric, and combined gas chromatography-mass spectrometric analysis of 78 toxic organic compounds has been determined. The toxic compounds are those routinely monitored in ambient air samples using Tenax collection and gas chromatography-mass spectrometric analysis. The Shannon information content of the binary encoded and full intensity mass spectra, of the gas chromatographic retention times, and of the combined gas chromatographic-mass spectrometric spectra of the 78 compounds has been calculated. The maximum binary information contents of the 35 channel gas chromatographic, 17 key channel mass spectral, and the 595 channel gas chromatographic-mass spectral methods were 6.4, 15.4, and < 21.8 bits, respectively. The 17 masses with the highest binary information content with regard to the 78 compounds were used with SIMCA pattern recognition to determine four classes among the 78 compounds. These included aromatics without chlorine substitution, chloroaromatics, bromoalkanes and alkenes, and chloroalkanes and alkenes.