猫群RBF神经网络诊断方法研究  

Research of Diagnosis Method Based on Cat Swarm RBF Neural Network

作　　者：马艳梅[1] MA Yanmei(Department of Information and Electrical Engineering,Huainan Vocational Technical College,Huainan 232001,China)

机构地区：[1]淮南职业技术学院信息与电气系,安徽淮南232001

出　　处：《新乡学院学报》2018年第3期29-31,共3页Journal of Xinxiang University

基　　金：安徽省自然科学基金项目(KJ2017A642)

摘　　要：神经网络的参数设置直接影响其诊断正确率。以猫群算法优化径向基函数神经网络参数,建立最优的RBF神经网络诊断模型,利用建立的诊断模型进行肺癌诊断实验分析。结果表明:所建猫群RBF神经网络诊断模型有较高的诊断正确率,克服了RBF神经网络参数确定的盲目性,对肺癌患者判断的准确率要高于一般的BP神经网络。Parameter setting for neural network directly affects the diagnosis performance.In this paper,RBF neural network parameters were optimized by cat swarm algorithm.The optimal RBF neural network model was established.It was used to diagnose early lung cancer and carry out experiment analysis.The test results showed that the presented model had higher diagnosis accuracy,overcoming the blindness of parameter setting for RBF.It was able to judge the patients with lung cancer more accurately and better than BP neural network diagnosis model.

关 键 词：猫群算法 RBF神经网络 诊断方法 参数优化 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]