STATISTICAL METHODS IN SIGNAL PROCESSING AND DISCRIMINATION

We deal with the classification of acoustic emission signals by means of Fuzzy Clustering (FC), Model-Based Clustering (MBC) and Support Vector Machines (SVM). These methods belong to a different group of classification techniques, e.g. the SVM is searching for optimal separating hyperplanes between clusters. The signals are compared by means of suitable parameters obtained directly from the signals and from nonmed frequency spectra such as phi-divergence distance measure as the additional attribute. We are concerned with resulting cluster comparisons and the selection of efficient classification parameters. We realize three experiments in the area of acoustic emission to test the proposed classification methods by means of laboratory data and also considering industrial data from the real life.