| 基于人工神经网络的特征压缩方法及应用 |
| |
| 引用本文： | 刘玮,乔新勇,安钢.基于人工神经网络的特征压缩方法及应用[J].装甲兵工程学院学报,2009,23(3):38-40,72. |
| |
| 作者姓名： | 刘玮 乔新勇 安钢 |
| |
| 作者单位： | 装甲兵工程学院机械工程系,北京,100072 |
| |
| 摘 要： | 分析了人工神经网络特征提取方法用于特征压缩的原理,研究了多层网络的隐层提取模型、Oja网络模型和基于Sanger算法的网络模型,以坦克柴油机燃油系统典型故障的特征处理为例,采用Sanger算法对特征参数进行了压缩,实现了特征空间的约简。该方法有利于简化分类器的设计,对于解决复杂设备的状态检测与故障诊断问题具有重要意义。
|
| 关 键 词： | 人工神经网络 特征压缩 柴油机 故障诊断 |
Feature Compression Methods and Applications of Artificial Neural Network |
| |
| LIU Wei,QIAO Xin-yong,AN Gang.Feature Compression Methods and Applications of Artificial Neural Network[J].Journal of Armored Force Engineering Institute,2009,23(3):38-40,72. |
| |
| Authors: | LIU Wei QIAO Xin-yong AN Gang |
| |
| Institution: | ( Department of Mechanical Engineering, Academy of Armored Force Engineering, Beijing 100072, China) |
| |
| Abstract: | This paper analyzes the theories of artificial neural network in feature compression, studies the hidden layer extraction model of multi-layer nets, Oja net model and Sanger model. It also introduces the feature processing method for the typical faults of tank engine, utilizes Sanger arithmetic to compress the calculated features, and realizes the reduction of feature space. This method contributes to simplifying the fault classifier, and has great importance in resolving the problems in state measurement and fault diagnosis. |
| |
| Keywords: | artificial neural network feature compression diesel engine fault diagnosis |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
