Application of support vector machine´s parameters selection in echo target recognition

Evolutionary algorithms for selecting support vector machine (SVM) parameter values which are based on genetic algorithm and particle swarm optimization algorithm are researched in this paper, these algorithms have been successfully applied to the real underwater echo target recognition. Experimental comparison and analysis show that the evolutionary algorithms can identify optimal or near optimal parameter settings more efficient and faster than the traditional “grid-research”. Performance of the evolutionary algorithms is demonstrated on the complex underwater echo target character recognition dataset.