Classification of Mushroom Substrate Odors Using a Portable Electronic Nose and Neural Networks

Abstract

An electronic nose was used in conjunction with a human panel and artificial neural networks topredict human assessments on a hedonic (pleasantness) scale when confronted with mushroomsubstrate odors. Both laboratory and field odor samples were collected and presented to both theelectronic nose and the human panelists. The human panelists were required to rate the odorsamples on a hedonic scale. Neural networks were developed to predict the odor pleasantnessbased on electronic nose readings. The results obtained from the developed neural networks showpromise in the utility of electronic nose evaluation of odors.