摘要
为了实现多传感器信息数据在神经网络中得到更好的运用,解决样本数据大导致网络复杂的问题。本文利用基于Rosetta数据分析平台对UCI数据库的机械故障数据集粗糙约简,剔除冗余特征以减少神经网络的输入节点,简化网络结构、提高运算速度。
In order to make the information from multiple sensors to get better use in the neural network.,and solve the problem of more samples will lead to more complex network.In this paper,the mechanical failure in the UCI database have been reducted based on Rosetta platform,removed redundant features to reduce the neural network input nodes,simplify network structure and improve the processing speed.
关键词
粗糙集
矿井风机
故障诊断
属性约简
Rough Sets Mine Fans
Failure Analysis
Attribute reduction
