Current trends in artificial nose research are strongly influenced by knowledge of biological olfactory systems. Insects have evolved over millions of years to detect and maneuver toward a food source or mate, or away from predators. The insect olfactory system is able to identify volatiles on a time scale that matches their ability to maneuver. Here, biological olfactory sense organs, insect antennae, have been exploited in a hybrid-device biosensor, demonstrating the ability to identify individual strands of odor in a plume passing over the sensor on a sub-second time scale. A portable system was designed to utilize the electrophysiological responses recorded from a sensor array composed of male or female antennae from four or eight different species of insects (a multi-channel electroantennogram, EAG). A computational analysis strategy that allows discrimination between odors in real time is described in detail. Following a training period, both semi-parametric and k-nearest neighbor (k-NN) classifiers with the ability to discard ambiguous responses are applied toward the classification of up to eight odors. EAG responses to individual strands in an odor plume are classified or discarded as ambiguous with a delay (sensor response to classification report) on the order of 1 s. The dependence of classification error rate on several parameters is described. Finally, the performance of the approach is compared to that of a minimal conditional risk classifier.