Two-dimensional correlation coefficient mapping in gas chromatography: Jet fuel classification for environmental analysis

We demonstrate the feasibility of using two-dimensional correlation coefficient mapping to classify gas chromatograms of environmental hazards. Correct identification and classification of the contaminants is the prerequisite for their appropriate treatment and containments. A data set consisting of 76 gas chromatograms of eight types of jet fuels, which are common sources of hydrocarbon contamination in ground water, is examined with two-dimensional statistical sample-sample correlation coefficients. Analysis demonstrates that jet fuel samples of the same type correlate strongly with each other but less significantly with other jet fuel classes. According to the magnitude of the correlation coefficients between each pair of the samples, jet fuel types of each sample in the data set can be assigned with an accuracy of 100% through a leave-one-out cross validation (LOOCV) procedure. Correlation coefficient mapping is thus a promising method to classify samples of environmental importance.