Abstract Lavine, B.K., Ward, A.J.I., Han Jian Hwa, Smith, R.-K. and Taylor, O.R., 1990. Taxonomy based on chemical constitution: differentiation of heavily Africanized honeybees from moderately Africanized honeybees. Chemometrics and Intelligent Laboratory Systems , 8: 239–243. Heavily Africanized honeybees can be differentiated from moderately Africanized honeybees based on differences in chemical constitution. The test data (i.e. training set) consisted of 70 chromatograms of hydrocarbon extracts obtained from the cuticle, exocrine and wax glands of heavily Africanized and moderately Africanized foragers. Each chromatogram contained 65 peaks corresponding to a set of standardized retention time windows. However, only 13 of these 65 gas chromatographic (GC) peaks were used for discriminant analysis. The feature selection process that was employed to select these peaks was based on a priori considerations, e.g. well defined baseline, reliable integration counts. A discriminant function was developed from this set of 13 GC peaks using the linear learning machine. The linear discriminant correctly classified every chromatogram in the training set and also achieved a high classification success rate for the prediction set which contained 118 chromatograms.
[1]
Thomas L. Isenhour,et al.
Some Chemical Applications of Machine Intelligence
,
1971
.
[3]
Using discriminant analysis to estimate linear mixing proportions
,
1987
.
[4]
L. L. Jackson,et al.
Incorporation of labelled dietary n-alkanes into cuticular lipids of the grasshopper Melanoplus sanguinipes
,
1973
.
[5]
African Bees: Potential Impact in the United States
,
1985
.
[6]
A. J. Stuper,et al.
Nonparametric feature selection in pattern recognition applied to chemical problems
,
1975
.
[7]
P. Jurs,et al.
Taxonomy based on chemical constitution: Differentiation of Africanized honey‐bees from European honey‐bees
,
1988
.