摘要
传统的特征选择方法没有很好地考虑数据的模式特性而导致性能下降.RelierF是较为有效的特征选择方法,但存在特征权值随样本波动和不能去除冗余特征的问题.对此,从数据本身的模式特性出发,提出了可信间隔的概念和基于可信间隔进行特征选择的方法.以氧化铝回转窑烧结过程数据为实验数据进行特征选择和烧结工况识别实验,结果表明,所提出的方法能去除冗余特征,有效地提高了识别率.
Traditional feature selection methods don't consider the pattern feature of data, which leads to the performance degradation. ReliefF is a more effective method for feature selection, but the weight of feature can fluctuate with samples and the method cannot remove redundant feature. Therefore, the concept of authentic distance and a new method used to feature selection are proposed based on the pattern feature of data, and the feature selection and experiments of sintering working conditions recognition are operated with the data of alumina.rotary kiln sintering process as experimental data. The experimental results show that the proposed method can effectively remove unrelated features and greatly improve recognition rate.
出处
《控制与决策》
EI
CSCD
北大核心
2011年第8期1229-1232,1238,共5页
Control and Decision
基金
国家863计划重点项目(2007AA041404)
国家自然科学基金重点项目(50834009)
关键词
可信间隔
特征选择
支持向量机
模式识别
authentic distance
feature selection
support vector machine
pattern recognition