MATLAB人工神经网络

《MATLAB人工神经网络》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。

1、ArtificialNeuralNetworksAbstract:TheArtificialNeuralNetwork(ANN)isafunctionalimitationofsimplifiedmodelofthebiologicalneuronsandtheirgoalistoconstructuseful‘computers’forreal-worldproblemsandreproduceintelligentdataevaluationtechniqueslikepatternreco

2、gnition,classificationandgeneralizationbyusingsimple,distributedandrobustprocessingunitscalledartificialneurons.Thispaperwillpresentasimpleapplicationoftheartificialneuralnetwork:process,designandperformanceanalysis.1.WorkingprocessofArtificialNeural

3、NetworksAnartificialneuronmodelsthedendritesofabiologicalneuronbyreceivingoneormoreinputs,thenapplyingappropriateweights(positiveornegativemultipliers)tothesignals,andpassestheweightedinputstothesoma.Thesomaorbodyofanartificialneuron,receivestheweigh

4、tedinputsandcomputesasingleoutputsignal(adiscreteorcontinuousfunctionofthesumoftheweightedinputs)byapplyingathresholdortransferfunction(Yoon,1989).Thelastsectionofeachartificialneuronisanalogoustotheaxonofabiologicalneuron,andpassestheoutputsignaltoo

5、neormoresubsequentartificialneuronsortotheoutsideworld.Everynon-trivialartificialneuralnetworkcontainsaninputlayerandanoutputlayer.Mostalsocontainoneormoreintermediateprocessinglayers.TypesofNeuralNetworks:Bothfeed-forwardandrecurrentnetworksareexamp

6、lesofsupervisedlearning.Inunsupervisedlearning,noexternalteacherisrequired.Thesystemself-organizestheinputdata,discoveringforitselftheregularitiesandcollectivepropertiesofthedata.Thesefeed-forwardnetworkshavetheabilitytolearn.Todoso,anartificialneura

7、lnetworkmustlearntoproduceadesiredoutputbymodifyingtheweightsfromitsinputs.Theprocessofhowthisisdoneissimple.2.ProblemsA.9trainingsamples,361testingsamples.B.9trainingsamples,361testingsamples.C.11*11trainingsamples,41*41testingsamples.3.DesigningWei

8、ghtedSumactivationfunctionerrorfunctionStep1:initializetheweightparametersandotherparametersdefaultpoints=50;%%隐含层节点数inputpoints=2;%%输入层节点数outputpoints=2;%%输出层节点数Testerror=zeros(1,100);%每个测试点的误差记录a=zeros(1,inputpoints);%输入层节点值y=zeros(1,outputpoints);