To stimulate the biochemical sensor system, an organic vapor sensing d
evice based on a sensor array for recognizing solvents with different
organic functional groups has been constructed. The array consists of
12 piezoelectric quartz crystals, each coated with an adsorptive mater
ial with unselective, but different response characteristics. The step
wise regression analysis shows that a subset of seven sensors is suffi
cient for the classification purpose. The principal component analysis
is applied to the original and autoscaled data. The results show that
autoscaling of the original data substantially improves the classific
ation of alcoholic, carbonyl, etherous, and esterous groups. The array
have been applied to classification of spirit and beverage samples. (
C) 1995 Academic Press, Inc.