1982 Volume 31 Issue 9 Pages 483-488
Application of cluster analysis which is a nonparametric and unsupervised pattern recognition method was studied to classify 62 kinds of crude oils in the world. These oils were produced in the Middle East, Indonesia, Africa, Mexico, Malaysia, Brunei, China, Australia, Burma, and USSR. Chromatograms of the oils were obtained by gel permeation chromatography with an ultraviolet absorption detector. Stationary and mobile phases were HSG-15 (polystyrene gel) and tetrahydrofuran, respectively. Sectional areas of each chromatogram were printed out with a data processor and ten normalized sectional areas were used as input data to perform cluster analysis. Similarity of profiles of the chromatograms was evaluated by Euclidean distance, and centroid method was used to, merge two clusters into a new cluster. Branching tree diagram was obtained from results of the cluster analysis. The proposed method was found to be effective for the classification of the crude oils.
