ISSN:
1612-1112
Keywords:
Gas chromatography
;
Neural networks
;
Classification
;
Pattern recognition
Source:
Springer Online Journal Archives 1860-2000
Topics:
Chemistry and Pharmacology
Notes:
Summary Artificial Neural Networks are proposed for the classification of vegetable oils submitted to gas chromatography analysis through their FAME profiles. Three important neural networks are evaluated here: the Hopfield model as a content-addressable memory (CAM), the Hamming model and the multi-layer perceptron. A brief description of these nets is first presented and further implemented for the classification of known vegetable oils. After the learning step, unknown samples are presented to the nets and the identification step performed. The performances of these nets are compared for the recognition of popular vegetable oils, including several edible oils.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1007/BF02269696
Permalink