论文部分内容阅读
针对传统神经网络模式识别中存在网络结构难于确定、过学习、收敛速度慢、易陷入局部极小值等不足及标准支持向量回归机中未考虑各样本重要性的差异问题,结合变压器油中多组分气体监测传感器阵列,将改进型支持向量回归机应用于气体传感器阵列信号模式识别中。实验结果表明,改进后的模式识别方法在预测精度和泛化能力上都较传统神经网络和标准支持向量回归模式识别方法有明显提高,有效地解决了多组分气体监测传感器的交叉敏感问题。
Aiming at the problems that traditional neural network pattern recognition is difficult to determine the network structure, over learning, slow convergence, easy to fall into the local minima and the standard support vector regression machine does not consider the importance of each sample difference, combined with the transformer oil Component gas monitoring sensor array, the improved support vector regression machine used in gas sensor array signal pattern recognition. The experimental results show that the improved pattern recognition method can obviously improve the prediction accuracy and generalization ability compared with the traditional neural network and the standard support vector regression pattern recognition method, and effectively solve the cross-sensitivity problem of the multi-component gas monitoring sensor.