基于支持向量聚类改进的雷达信号分选方法

在复杂的电磁信号环境下,针对使用支持向量聚类算法进行信号分选时,聚类结果易受核函数参数和惩罚因子的影响,及计算复杂度高和准确率低的问题,提出了一种将支持向量聚类与集对分析相结合的雷达信号分选方法。该方法使用集对分析先对支持向量进行聚类,再用所得的聚类结果对剩余数据进行聚类。仿真结果表明该方法在含有一定数量的噪声信号环境下,能获得较好的分选效果,不仅耗时短,而且还具有较高识别率。

Improved Radar Signal Sorting Method Based on Support Vector Clustering

Under the complex electromagnetic signal environment, since support vector clustering algorithm for signal separation has a high complexity and poor accuracy, the results of which are susceptible to the influence of kernel function parameter and penalty factor. A radar signal sor-ting method is introduced combining support vector clustering with set pair analysis, which firstly uses set pair analysis method to presort support vector, and then uses the result to process the re-maining data. Simulation results show that this method can achieve a good separation effect un-der the environment of containing a certain amount of noise signal, not only does it require less time, but also has a higher rate of identification.