基于贝叶斯网络分类器的雷达辐射源识别方法

雷达辐射源识别是电子对抗中的重要组成部分.贝叶斯网络分类器建立在坚实的理论基础之上,具有较为优秀的分类性能,而且能够有效地处理不确定性问题,重点研究了如何利用贝叶斯网络分类器进行雷达辐射源识别,并通过仿真实验对朴素贝叶斯分类器及其扩展方法进行了分析比较.实验结果表明,与基于概率近似准则的方法相比,基于分类准确率提高准则的扩展树生成方法具有更为优秀的分类性能.

Recognition Method of Radar Emitters based on Bayesian Network Classifiers

The recognition of radar emitters is an important component of Electronic Warfare.Bayesian Network Classifiers has a solid grounding in statistics.They have surprisingly good classification performance and they could deal with uncertainty effectively.This paper focuses on the recognition of radar emitters using Bayesian Network Classfiers.Diferent models such as Naive Bayes Classifier and its augmenting appraches are analyzed and compared.The experimentral result shows that classification-based augmenting approaches are more effective than that of distribution-based.