基于改进LSSVM的短波收信天线智能诊断研究  

Intelligent Diagnosis Research of Short-wave Receiving Antenna Based on Improved LSSVM

作　　者：向奕雪 陈斌[1] 罗勇[1] XIANG Yixue;CHEN Bin;LUO Yong(Naval University of Engineering,Wuhan 430033)

机构地区：[1]海军工程大学

出　　处：《计算机与数字工程》2019年第6期1331-1337,共7页Computer & Digital Engineering

摘　　要：长期以来短波收信天线系统一直缺少智能化的自动监测技术与手段,难以对设备的收信效果实时评估和预警。为了减轻基层通信保障人员的负担,同时为短波收信天线的健康状态分析与评估、设备的维修保障提供切实可靠的数据支持与技术支撑。论文建立基于果蝇优化算法的最小二乘支持向量机(LSSVM)分类器,实现了短波收信天线的智能诊断。并针对最小二乘支持向量机的稀疏化问题,从代表性样本与边界样本两个方面综合分析,提出了一种基于KFCM聚类的LSSVM训练样本预选取算法。UCI数据集的实验结果表明:该样本预选取算法充分利用了KFCM聚类的优点,有效地提取了启发信息更为丰富的样本并去除训练集中的冗余信息,获得了更优性能的分类器。For a long time,the short wave receiving antenna system has been lack of intelligent automatic monitoring technology and means,and it is difficult to realize real-time evaluation and early warning of the receiving effect of the equipment. In order to reduce the burden of the basic communication security personnel and provide reliable data support and technical support for the analysis and evaluation of the health condition of short wave receiving antennas and the maintenance and guarantee of the equipment,in this paper,a least squares support vector machine classifier based on fruit fly optimization algorithm is established,and the intelligent diagnosis of short wave receiving antenna is realized. To solve the sparseness of the least squares support vector machines,considering the two aspects of the representative samples and the boundary samples,a LSSVM training algorithm based on KFCM clustering algorithm is proposed. The UCI experimental results show that this method takes advantages of KFCM clustering to extract samples with more abundant heuristic information and remove redundant information effectively,a classifier with better performance is achieved.

关 键 词：智能诊断 核模糊C均值聚类 最小二乘支持向量机 果蝇优化算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]