局域自适应子波高斯神经网络综合分类系统

本文提出了一种用于船舶噪声分类的局域自适应子波高斯神经网络综合分类系统。该系统融合了两种特征提取和分类方法,即自适应子波神经网络和自适应高斯神经网络分类器,并利用网络局域化使得系统具有追加学习的能力。通过对实际的三类船舶噪声进行分类识别,结果令人满意,证明了该方法的优越性和工程应用前景。

A LOCAL ADAPTIVE WAVELET AND GAUSS NEURAL NETWORK SYNTHESIS CLASSIFICATION SYSTEM

In this paper, an efficient engineering classification of ship noises based on a local adaptive wavelet and Gauss neural network synthesis classification system is presented. The classification systems combine two methods of feature extraction and classification, which are adaptive wavelet neural network and adaptive Gauss neural network. It is capable of learning new types of signals and not destroying the learned network. The classification system is used to extract automatically feature from and classify for noises radiated from actual three types of ships. The classified results are encouraging, and this method is proved to be superior and efficient engineering application in the future.