摘要
主要探索一种从庞杂数据中挖掘有用信息的方法。首先介绍了粗糙集的基本理论与计算近似精度的方法 ,简述了粗糙集理论的特点及与模糊集理论、证据理论的区别与联系 ,然后将经过预处理的发动机振动信号进行实数离散 ,运用粗糙集的下近似、上近似及粗糙逼近理论 ,计算属性等价类对决策等价类的逼近精度。计算结果表明 ,采用等频率和等量间隔相结合的方法离散实数能保留数据中良好的自然分类特性 ,采用粗糙集的近似逼近理论能有效地提取出发动机故障特征。
This paper presents a way of mining useful information from a mass of data.At first,the basic theory of rough set and the method of calculating approximative extent are introduced,and the characters of tough set and difference and relation between rough set,fuzzy set and evidence theory are described.Then,the pretreated engine’s vibration signal is real-number-scattered,and the lower,upper approximation and rough approaching theory of rough set are applied to calculate the approaching extent of the attribute indiscernibity class to the decision-making indiscernibity class.The results show that applying the method combining equal-frequency and equal-number interval can scatter real-number under the condition of keeping the good natural classing traits in data;and applying the approximative approaching theory of rough set can extract the features from engine’s vibration signals effectively
出处
《振动．测试与诊断》
EI
CSCD
2004年第4期262-265,共4页
Journal of Vibration,Measurement & Diagnosis
基金
国防预研基金资助项目 (编号 :0 4 0 10 10 30 10 8)
关键词
粗糙集
发动机
振动信号
故障特征提取
rough set engine vibration signal fault feature extraction
