Identification of aromas from wine using quartz-resonator gas sensors in conjuction with neural-network analysis

Abstract Transient response curves for aromas that are generated by vapourizing wine by a heater are observed using only an epoxy-coated quartz-resonator gas sensor. The pattern-recognition analysis using principal-component or neural network analysis is carried out using nine parameters that characterize the transient response curves. The recognition probability of the neural-network analysis is 100% for three kinds of wine, red, white and rose´.