Prediction of hedonic tone using an electronic nose and artificial neural networks

An electronic nose was used in conjunction with a human panel and artificial neural networks to predict human assessments on a hedonic (pleasantness) scale of various odors, including fragrances, spent and phase I mushroom substrate odors. After calculating the average rating from the human panel for each odor sample, the hedonic rating was matched with the 32 electronic nose sensor readings from the corresponding sample. The data were then used to develop and train artificial neural networks. Several neural networks were developed and tested by comparing predicted hedonic tone from the neural networks with the average hedonic tone assessed by the human panel. The general regression neural network using 11 e-nose sensors as input gave the best test set results, yielding an r² value of 0.92, and a root mean error of 0.06 hedonic tone units. The genetic strategy neural network gave the best validation results, with an r² value of 0.67 and a root mean error of 1.11 pleasantness units. These results provide evidence that an electronic nose could potentially be used to assess the quality of odors in a similar manner to a human panel.