A method for efficient classification of microphones based on expert knowledge and computational intelligence

Abstract

Classification of microphones as photoacoustic detectors is important part of procedure of photoacoustic measurements calibration. The requirements of photoacoustic experiment are accuracy, precision, reliability and work in real time in order to be competitive measurement technique. According to current state, real time is still a problem. This paper suggests improvement of classification method currently in use by dimensionality reduction of input vector considered in the data preprocessing, having consequence in significant simplification of measurements and thus notable decrease of measurement time, so reaching real time calibration procedure. By applying the method presented in the article the number of measurement points will be one, two or three depending on its position on frequency axes which is extremely smaller number than commonly accepted (usually 70 - 80 for the frequency range 20 Hz - 20 000 Hz). The method is based on computational intelligence algorithms and expert knowledge.