欢迎来到天天文库
浏览记录
ID:53028669
大小:810.68 KB
页数:4页
时间:2020-04-14
《基于RBF神经网络的老年痴呆症智能诊断研究-论文.pdf》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、基于RBF神网络的老年痴导瘟智雷g诊断研究张会敏①叶明全罗永钱②孟婷玮①陈碉珠③摘要为验证单RBF神经网络更适用于老年痴呆症的预测诊断,通过仿真实验将单BP神经网络、单RBF神经网络、遗传算法优化BP神经网络及遗传算法优化RBF神经网络分别应用于老年痴呆症的预测诊断,建立这四种网络模型,并对四种网络模型的预测结果进行分析比较。仿真实验在Matlab软件平台上进行。结果表明:在老年痴呆症的预测诊断中,单RBF神经网络比单BP神经网络预测结果更好,建模时间更短。此外,单RBF神经网络与遗传算法优化的BP神经网络预测结果相同,但单RBF神经网络建模较为简单,预测结果更为稳定。而遗传
2、算法对RBF神经网络优化作用不明显。因此,单RBF神经网络更适用于老年痴呆症的预测诊断,实际应用时可以此结论作为理论指导。关键词遗传算法BP神经网络RBF~*经网络老年痴呆症预测数据挖掘Doi:lO.3969~.issn.1673—7571.2015.06.011【中图分类号】R741;TP391【文献标识码】AStudyonIntelligentDiagnosisofSenileDementiaBasedonRBFNeuralNetwork/ZHANGHui—min,YEMing—quan,LUOYong—qian,etal//ChinaDistalMedicine,一20
3、1510(6):38tO41AbstractInordertOverifysingleRBFneuralnetworkismoresuitableforthepredictivediagnosisofseniledementia,throughthesimulationexperiment,asingleBPneuralnetwork,asingleRBFneurNnetwork,ageneticalgorithmtOoptimizeBPneuralnetworkandageneticalgorithmtooptimizeRBFneuralnetworkareusedtopr
4、edictseniledementia,establishingofthesefourkindsofnetworkmodel,thenanalyzingandcomparingtheforecastedresultsofthesefourkindsofnetworkmode1.ThesimulationexperimentswerecarriedoutontheplatformofMatlabsoftware,theresultsshowthat:inthepredictivediagnosisofseniledementia,thesingleRBFneur~network
5、predictiveresultsishigherthanthesingleBPneuralnetwork,andthemodelingtimeisshorterFurthermore,thepredictionresultsofthesingleRBFneuralnetworkisasthesameasthegeneticalgorithmtOoptimizeBPneuralnetwork,butthesingleRBFneuralnetworkmodelisrelativelysimple,andthepredictionresultsaremorestable.Ther
6、efore,diagnosisandpredictionofthesingleRBFneurNnetworkismoresuitableforseniledementia,andthisconclusioncanbeusedasatheoreticalguidetotheactualapplication.Keywordsgeneticalgorithm,BPneur~network,RBFneuralnetwork,dementiadiseaseprediction,dataminingFundprojectNational—levelCollegeStudents’Inn
7、ovativeTrainingProgram(No.201310368()27);Provincial—levelCollegeStudents’InnovativeTrainingProgram(No.AH201310368027,AH201410368072);KeyProjectofProvincial—levelNaturalScienceResearchofCollegesandUniversitiesinAnhuiProvince(No.KJ2014A266)Correspondingaut
此文档下载收益归作者所有