Classification of multicomponent analytical data of olive oils using different neural networks

摘要: Abstract A comparison of classification abilities two different neural network methods, namely, back-propagation errors and Kohonen learning is made discussed. The performed on a set 572 Italian olive oils the basis an analysis eight fatty acids. methods carried out by architectures for each strategy separately. It was found that applied problem superior to errors. Additionally, levels weights in can be exploited give more detailed information about separation ability individual variable, i.e. acid our case.