基于小波熵理论的决策树信号分类识别算法  

Decision tree signal classification and recognition algorithm based on wavelet entropy theory

作　　者：李靖超[1] 钱迪 董春蕾 LI Jingchao;QIAN Di;DONG Chunlei(School of Electronic Information Engineering,Shanghai Dianji University,Shanghai 201306,China)

机构地区：[1]上海电机学院电子信息学院,上海201306

出　　处：《上海电机学院学报》2019年第2期100-103,共4页Journal of Shanghai Dianji University

基　　金：国家自然科学基金青年基金资助项目(61603239)

摘　　要：在信息理论的基础上有效提取信号的熵特征是信号特征提取的方式之一。在熵理论的基础上,提出了基于小波熵的特征提取算法,实现对频率调制信号的类内识别。设计决策树分类器,对不同信噪比下的4种待识别通信信号进行分类。仿真结果表明:基于小波熵理论的特征提取算法能够有效提取不同信号特征,并利用决策树分类器,通过设定阈值进行分类,识别率高,有利于信号的类内识别,具有广泛的应用价值。One of the methods of signal feature extraction is to effectively extract entropy features of the signal.With the entropy theory,the feature extraction method based on the wavelet entropy is proposed to realize the interclass recognition of the frequency modulated signal.Then a decision tree classifier is designed to classify four kinds of communication signals to be identified under different signals(signal noise ratio environment).The simulation results show that the feature extraction method based on the wavelet entropy theory can effectively extract different signal features,and the decision tree classifier can be utilized to classify signals through threshold setting,which has a high recognition rate.Therefore,it is conducive to interclass recognition of signals and has high application value.

关 键 词：信号识别 小波熵特征提取 决策树分类器 

分 类 号：TN91[电子电信—通信与信息系统]