Record Display for the EPA National Library Catalog

RECORD NUMBER: 3 OF 26

OLS Field Name OLS Field Data
Main Title Comparison of SIMCA Pattern Recognition and Library Search Identification of Hazardous Compounds from Mass Spectra.
Author Sarker, M. ; Glen, W. G. ; Yin, L. B. ; Dunn, W. J. ; Scott, D. R. ;
CORP Author Illinois Univ. at Chicago. ;Midwest Research Inst., Kansas City, MO.;Environmental Protection Agency, Research Triangle Park, NC. Atmospheric Research and Exposure Assessment Lab.
Publisher c31 Oct 90
Year Published 1990
Report Number EPA-R-814363; EPA/600/J-92/294;
Stock Number PB92-227362
Additional Subjects Hazardous materials ; Pattern recognition ; Mass spectroscopy ; Solid wastes ; Air pollution ; Polychlorinated biphenyls ; Halohydrocarbons ; Hydrocarbons ; Statistical analysis ; Reprints ; SIMCA methods
Holdings
Library Call Number Additional Info Location Last
Modified
Checkout
Status
NTIS  PB92-227362 Most EPA libraries have a fiche copy filed under the call number shown. Check with individual libraries about paper copy. NTIS 06/01/1993
Collation 12p
Abstract
SIMCA pattern recognition methods have been applied to mass spectral data from a target list of hazardous chemicals. A scheme has been proposed for classification and identification of five classes of compounds including aromatics, chlorocarbons, bromocarbons, hydrocarbons and polychlorinated biphenyls (PCBs). In addition, partial least squares regression has been used to predict the number of chlorine atoms present in the PCBs. A training set of 429 compounds was used. Classification models were derived by using principal components analysis of the autocorrelation transformed spectra and were-applied to two gas chromatography-mass spectrometry ambient air field samples and one hazardous waste sample. An optimal number of three principal components was determined. For identification the mass spectra of the unknowns were compared with the training set mass spectra in the predicted class. The overall classification and identification rates for training spectra were 89% and for the unknowns were 92% and 82%, respectively. Identification results of the SIMCA scheme for an ambient air field sample also was compared to that of probability based matching (PBM). Correct identification for the spectra in the sample by SIMCA was 71/84 (85%) as opposed 35/84 (42%) by PBM.