A method of analysis and automatic classification of the biostochastic signal is reported. It is based on the amplitude-domain characteristics of the signal. Two deterministic feature vectors are defined and then computed using a statistical algorithm. The classification is performed by matching these two features of the unknown signal with those of the known templates. The decision criterion has been determined and confirmed experimentally. This work suggests that the method of analysis has potential as a microcomputer-based diagnostic assist device for EEG, EMG, and auscultatory signals.<<ETX>>