摘要
雷达信号分选的容差问题一直是雷达对抗情报处理中的难题,而支持向量聚类法(SVC)是一种能够有效避免容差问题的聚类方法,但现有参数搜索方法不能快速准确地确定SVC最优聚类结构,从而限制了支持向量聚类法的广泛应用。针对这一问题,提出了一种可以自动选择参数的SVC聚类方法。它通过采用一种综合的参数搜索方法,自动选择惩罚因子和核函数宽度两个参数,从而确定最优的聚类结构。仿真实验表明,此方法可在较少的迭代次数下获得最优的聚类结构,提高了雷达信号的分选正确率。
The problem of tolerance of radar signal sorting is always complicated in radar countermeasure. Support Vector Clustering (SVC) is a clustering approach which can avoid the problem of toler- ance effectively. But it restricts the practical application of the SVC as the existing parameters search method cannot get the optimal clustering model. A new SVC algorithm with automatic parameters selection is proposed based on the analysis of the problem. The algorithm can select the regularization parameter and parameter of the kernel function, which uses a integrated parameter search method to obtain the optimal clustering model. Experiments show that the proposed approach can get better clustering model with the less iterative operation, which can improve the sorting ratio.
出处
《电子信息对抗技术》
2011年第2期15-20,73,共7页
Electronic Information Warfare Technology
关键词
支持向量聚类
信号分选
最佳聚类模型
雷达信号
support vector clustering
signal sorting
optimal clustering model
radar signal