Classification of essential mint oils of different geographic origin by applying pattern recognition methods to gas chromatographic data

Abstract Pattern recognition techniques were applied to the classification of essential mint oils using their gas chromatographic data. A study has been completed on 59 samples of three different geographic origins: U.S.A., France and Italy. Principal component analysis provided new variables for an effective classification of the samples according to the different areas of origin. The pattern of the result was further analysed by means of a fuzzy clustering algorithm which permitted the quantification of the differences between the three classes. The chemical information contained in the gas chromatographic data was sufficient to characterize the geographic origin of the samples.