Computational intelligence based method for efficient classification of microphones

Abstract

This paper presents a new, improved classification method of microphones as photoacoustic detectors by applying computational intelligence algorithms combined with expert knowledge. The classification method is part of the procedure of calibration of model dependent diagnostic technics. Novelty is dimensionality reduction to one, two or three measurement points depending on the point position on the frequency axis and it was considered in the data preprocessing. It has been proven that the presented classification method is accurate, reliable and needs less time than standard classification procedures, enabling fast and the precise processing of photoacoustic measurement data. Special achievement refers to the experimental procedure, regarding reduction in the number of measurement points needed for classification in comparison to the number usually used in the standard measurement procedure of the photoacoustic experiment.