Matlab的神經(jīng)網(wǎng)絡(luò)工具箱入門(mén)

Matlab的神經(jīng)網(wǎng)絡(luò)工具箱入門(mén)在commandwindow中鍵入helpnnet>>helpnnetNeuralNetworkToolboxVersion7.0（R2010b）03-Aug-2010神經(jīng)網(wǎng)絡(luò)工具箱版本7.0（R2010b）03八月，2010圖形用戶(hù)界面功能。nnstart-神經(jīng)網(wǎng)絡(luò)啟動(dòng)GUInctool-神經(jīng)網(wǎng)絡(luò)分類(lèi)工具nftool-神經(jīng)網(wǎng)絡(luò)的擬合工具nntraintool-神經(jīng)網(wǎng)絡(luò)的訓(xùn)練工具nprtool-神經(jīng)網(wǎng)絡(luò)模式識(shí)別工具ntstool-NFTool神經(jīng)網(wǎng)絡(luò)時(shí)間序列的工具nntool-神經(jīng)網(wǎng)絡(luò)工具箱的圖形用戶(hù)界面。查看-查看一個(gè)神經(jīng)網(wǎng)絡(luò)。網(wǎng)絡(luò)的建立功能。cascadeforwardnet-串級(jí)，前饋神經(jīng)網(wǎng)絡(luò)。competlayer-競(jìng)爭(zhēng)神經(jīng)層。distdelaynet-分布時(shí)滯的神經(jīng)網(wǎng)絡(luò)。elmannet-Elman神經(jīng)網(wǎng)絡(luò)。feedforwardnet-前饋神經(jīng)網(wǎng)絡(luò)。fitnet-函數(shù)擬合神經(jīng)網(wǎng)絡(luò)。layrecnet-分層遞歸神經(jīng)網(wǎng)絡(luò)。linearlayer-線(xiàn)性神經(jīng)層。lvqnet-學(xué)習(xí)矢量量化（LVQ）神經(jīng)網(wǎng)絡(luò)。narnet-非線(xiàn)性自結(jié)合的時(shí)間序列網(wǎng)絡(luò)。narxnet-非線(xiàn)性自結(jié)合的時(shí)間序列與外部輸入網(wǎng)絡(luò)newgrnn-設(shè)計(jì)一個(gè)廣義回歸神經(jīng)網(wǎng)絡(luò)。newhop-建立經(jīng)常性的Hopfield網(wǎng)絡(luò)。newlind-設(shè)計(jì)一個(gè)線(xiàn)性層。newpnn-設(shè)計(jì)概率神經(jīng)網(wǎng)絡(luò)。newrb-徑向基網(wǎng)絡(luò)設(shè)計(jì)。newrbe-設(shè)計(jì)一個(gè)確切的徑向基網(wǎng)絡(luò)。patternnet-神經(jīng)網(wǎng)絡(luò)模式識(shí)別。感知-感知。selforgmap-自組織特征映射。timedelaynet-時(shí)滯神經(jīng)網(wǎng)絡(luò)。利用網(wǎng)絡(luò)。網(wǎng)絡(luò)-創(chuàng)建一個(gè)自定義神經(jīng)網(wǎng)絡(luò)。SIM卡-模擬一個(gè)神經(jīng)網(wǎng)絡(luò)。初始化-初始化一個(gè)神經(jīng)網(wǎng)絡(luò)。適應(yīng)-允許一個(gè)神經(jīng)網(wǎng)絡(luò)來(lái)適應(yīng)?；疖?chē)-火車(chē)的神經(jīng)網(wǎng)絡(luò)。DISP鍵-顯示一個(gè)神經(jīng)網(wǎng)絡(luò)的屬性。顯示-顯示的名稱(chēng)和神經(jīng)網(wǎng)絡(luò)屬性adddelay-添加延遲神經(jīng)網(wǎng)絡(luò)的反應(yīng)。closeloop-神經(jīng)網(wǎng)絡(luò)的開(kāi)放反饋轉(zhuǎn)換到關(guān)閉反饋回路。formwb-表格偏見(jiàn)和成單個(gè)向量的權(quán)重。getwb-將它作為一個(gè)單一向量中的所有網(wǎng)絡(luò)權(quán)值和偏差noloop-刪除神經(jīng)網(wǎng)絡(luò)的開(kāi)放和關(guān)閉反饋回路。開(kāi)環(huán)-轉(zhuǎn)換神經(jīng)網(wǎng)絡(luò)反饋，打開(kāi)封閉的反饋循環(huán)。removedelay-刪除延遲神經(jīng)網(wǎng)絡(luò)的反應(yīng)。separatewb-獨(dú)立的偏見(jiàn)和重量/偏置向量的權(quán)重。setwb-將所有與單個(gè)矢量網(wǎng)絡(luò)權(quán)值和偏差。Simulink的支持。gensim-生成Simulink模塊來(lái)模擬神經(jīng)網(wǎng)絡(luò)。setsiminit-集神經(jīng)網(wǎng)絡(luò)的Simulink模塊的初始條件getsiminit-獲取神經(jīng)網(wǎng)絡(luò)Simulink模塊的初始條件神經(jīng)元-神經(jīng)網(wǎng)絡(luò)Simulink的模塊庫(kù)。培訓(xùn)職能。trainb-批具有重量與偏見(jiàn)學(xué)習(xí)規(guī)則的培訓(xùn)。trainbfg-的BFGS擬牛頓倒傳遞。trainbr-貝葉斯規(guī)則的BP算法。trainbu-與重量與偏見(jiàn)一批無(wú)監(jiān)督學(xué)習(xí)規(guī)則的培訓(xùn)。trainbuwb-與體重?zé)o監(jiān)督學(xué)習(xí)規(guī)則與偏見(jiàn)一批培訓(xùn)。trainc-循環(huán)順序重量/偏見(jiàn)的培訓(xùn)。traincgb-共軛鮑威爾比爾重新啟動(dòng)梯度反向傳播。traincgf-共軛弗萊徹-里夫斯更新梯度反向傳播。traincgp-共軛波拉克-Ribiere更新梯度反向傳播。traingd-梯度下降反向傳播。traingda-具有自適應(yīng)LR的反向傳播梯度下降。traingdm-與動(dòng)量梯度下降。traingdx-梯度下降瓦特/慣性與自適應(yīng)LR的反向傳播。trainlm-采用Levenberg-馬奎德倒傳遞。trainoss-一步割線(xiàn)倒傳遞。trainr-隨機(jī)重量/偏見(jiàn)的培訓(xùn)。trainrp-RPROP反向傳播。trainru-無(wú)監(jiān)督隨機(jī)重量/偏見(jiàn)的培訓(xùn)?；疖?chē)-順序重量/偏見(jiàn)的培訓(xùn)。trainscg-規(guī)?；曹椞荻菳P算法。繪圖功能。plotconfusion-圖分類(lèi)混淆矩陣。ploterrcorr-誤差自相關(guān)時(shí)間序列圖。ploterrhist-繪制誤差直方圖。plotfit-繪圖功能適合。plotinerrcorr-圖輸入錯(cuò)誤的時(shí)間序列的互相關(guān)。plotperform-小區(qū)網(wǎng)絡(luò)性能。

plotregression-線(xiàn)性回歸情節(jié)。plotresponse-動(dòng)態(tài)網(wǎng)絡(luò)圖的時(shí)間序列響應(yīng)。plotroc-繪制受試者工作特征。plotsomhits-小區(qū)自組織圖來(lái)樣打。plotsomnc-小區(qū)自組織映射鄰居的連接。plotsomnd-小區(qū)自組織映射鄰居的距離。plotsomplanes-小區(qū)自組織映射重量的飛機(jī)。plotsompos-小區(qū)自組織映射重量立場(chǎng)。plotsomtop-小區(qū)自組織映射的拓?fù)浣Y(jié)構(gòu)。plottrainstate-情節(jié)訓(xùn)練狀態(tài)值。plotwb-圖寒春重量和偏差值圖。列出其他神經(jīng)網(wǎng)絡(luò)實(shí)現(xiàn)的功能。nnadapt-適應(yīng)職能。nnderivati??ve-衍生功能。nndistance-距離函數(shù)。nndivision-除功能。nninitlayer-初始化層功能。nninitnetwork-初始化網(wǎng)絡(luò)功能。nninitweight-初始化權(quán)函數(shù)。nnlearn-學(xué)習(xí)功能。nnnetinput-凈輸入功能。nnperformance-性能的功能。nnprocess-處理功能。nnsearch-線(xiàn)搜索功能。nntopology-拓?fù)浣Y(jié)構(gòu)的功能。nntransfer-傳遞函數(shù)。nnweight-重量的功能。示威，數(shù)據(jù)集和其他資源nndemos-神經(jīng)網(wǎng)絡(luò)工具箱的示威。nndatasets-神經(jīng)網(wǎng)絡(luò)工具箱的數(shù)據(jù)集。nntextdemos-神經(jīng)網(wǎng)絡(luò)設(shè)計(jì)教科書(shū)的示威。nntextbook-神經(jīng)網(wǎng)絡(luò)設(shè)計(jì)教科書(shū)的資訊。調(diào)出圖形用戶(hù)界面如下用他給出的命令nntool（同樣是在commandwindow鍵入）調(diào)出圖形用戶(hù)界面如下■InputData:Import...New...⑥HelpQClose>>helpnnetNeuralNetworkToolbox神經(jīng)網(wǎng)絡(luò)工具箱Version8.1(R2013b)08-Aug-20132013的8.1版Graphicaluserinterfacefunctions.圖形用戶(hù)界面函數(shù)nnstart-NeuralNetworkStartGUI神經(jīng)網(wǎng)絡(luò)啟動(dòng)圖形用戶(hù)界面GUInctool-Neuralnetworkclassificationtool神經(jīng)網(wǎng)絡(luò)分類(lèi)工具nftool-NeuralNetworkFittingTool神經(jīng)網(wǎng)絡(luò)擬合工具nntraintool-Neuralnetworktrainingtool神經(jīng)網(wǎng)絡(luò)訓(xùn)練工具nprtool-Neuralnetworkpatternrecognitiontool神經(jīng)網(wǎng)絡(luò)模式識(shí)別工具ntstool-NFToolNeuralNetworkTimeSeriesToolNFTool神經(jīng)網(wǎng)絡(luò)時(shí)間序列工具nntool-NeuralNetworkToolboxgraphicaluserinterface.神經(jīng)網(wǎng)絡(luò)工具箱圖形用戶(hù)界面view-Viewaneuralnetwork.查看一個(gè)神經(jīng)網(wǎng)絡(luò)Networkcreationfunctions.網(wǎng)絡(luò)生成函數(shù)cascadeforwardnet-Cascade-forwardneuralnetwork.級(jí)聯(lián)神經(jīng)網(wǎng)絡(luò)competlayer-Competitiveneurallayer.競(jìng)爭(zhēng)神經(jīng)層distdelaynet-Distributeddelayneuralnetwork.分布式延遲神經(jīng)網(wǎng)絡(luò)elmannet-Elmanneuralnetwork.Elman神經(jīng)網(wǎng)絡(luò)feedforwardnet-Feed-forwardneuralnetwork.前饋神經(jīng)網(wǎng)絡(luò)fitnet-Functionfittingneuralnetwork.函數(shù)擬合神經(jīng)網(wǎng)絡(luò)layrecnet-Layeredrecurrentneuralnetwork.分層遞歸神經(jīng)網(wǎng)絡(luò)linearlayer-Linearneurallayer.線(xiàn)性神經(jīng)層lvqnet-Learningvectorquantization(LVQ)neuralnetwork.學(xué)習(xí)向量量化(LVQ)神經(jīng)網(wǎng)絡(luò)

narnet間序列網(wǎng)絡(luò)narxnet-Nonlinearauto-associativetime-seriesnetwork.非線(xiàn)性自動(dòng)組合時(shí)-Nonlinearauto-associativetime-seriesnetworkwithexternalinput.具有外部輸入的非線(xiàn)性自動(dòng)組合時(shí)間序列網(wǎng)絡(luò)newgrnn-Designageneralizedregressionneuralnetwork.設(shè)計(jì)一個(gè)廣義回歸神經(jīng)網(wǎng)絡(luò)newhop網(wǎng)絡(luò)newlind-CreateaHopfieldrecurrentnetwork.創(chuàng)建一個(gè)Hopfield復(fù)發(fā)性-Designalinearlayer.設(shè)計(jì)一個(gè)線(xiàn)性層newpnnnewrbnewrbepatternnetperceptronselforgmap-Designaprobabilisticneuralnetwork.設(shè)計(jì)一個(gè)概率神經(jīng)網(wǎng)絡(luò)-Designaradialbasisnetwork.設(shè)計(jì)一個(gè)徑向基網(wǎng)絡(luò)-Designanexactradialbasisnetwork.設(shè)計(jì)一個(gè)精確的徑向基網(wǎng)絡(luò)-Patternrecognitionneuralnetwork.模式識(shí)別神經(jīng)網(wǎng)絡(luò)-Perceptron.-Self-organizingmap.自組織映射timedelaynet-Time-delayneuralnetwork.時(shí)間延遲神經(jīng)網(wǎng)絡(luò)Usingnetworks.網(wǎng)絡(luò)使用networksiminitadapttraindispdisplay名字和屬性-Createacustomneuralnetwork.創(chuàng)建一個(gè)定制的神經(jīng)網(wǎng)絡(luò)-Simulateaneuralnetwork.模擬神經(jīng)網(wǎng)絡(luò)-Initializeaneuralnetwork.初始化一個(gè)神經(jīng)網(wǎng)絡(luò)-Allowaneuralnetworktoadapt.神經(jīng)網(wǎng)絡(luò)的適應(yīng)-Trainaneuralnetwork.訓(xùn)練一個(gè)神經(jīng)網(wǎng)絡(luò)-Displayaneuralnetwork'sproperties.顯示一個(gè)神經(jīng)網(wǎng)絡(luò)的屬性-Displaythenameandpropertiesofaneuralnetwork顯示一個(gè)神經(jīng)網(wǎng)絡(luò)的adddelaycloseloop-Addadelaytoaneuralnetwork'sresponse.加一個(gè)延時(shí)到神經(jīng)網(wǎng)絡(luò)響應(yīng)-Convertneuralnetworkopenfeedbacktoclosedfeedbackloops.轉(zhuǎn)變神經(jīng)網(wǎng)絡(luò)打開(kāi)反饋到關(guān)閉反饋的回路formwbgetwb-Formbiasandweightsintosinglevector.使偏差和權(quán)重成為單一向量-Getallnetworkweightandbiasvaluesasasinglevector.獲得全部網(wǎng)絡(luò)權(quán)重和偏差當(dāng)作單一向量noloop-Removeneuralnetworkopenandclosedfeedbackloops.移去神經(jīng)網(wǎng)絡(luò)打開(kāi)和關(guān)閉反饋回路openloop-Convertneuralnetworkclosedfeedbacktoopenfeedbackloops.轉(zhuǎn)變神經(jīng)網(wǎng)絡(luò)關(guān)閉反饋到打開(kāi)反饋的回路removedelay-Removeadelaytoaneuralnetwork'sresponse.為神經(jīng)網(wǎng)絡(luò)反應(yīng)移去一個(gè)延遲separatewb-Separatebiasesandweightsfromaweight/biasvector.從一個(gè)權(quán)重/偏差向量分離偏差和權(quán)重setwb-Setallnetworkweightandbiasvalueswithasinglevector.用一個(gè)單一向量設(shè)置全部網(wǎng)絡(luò)權(quán)重和偏差值Simulinksupport.仿真支持gensim-GenerateaSimulinkblocktosimulateaneuralnetwork.生成Simulink模塊來(lái)模擬神經(jīng)網(wǎng)絡(luò)setsiminit-SetneuralnetworkSimulinkblockinitialconditions設(shè)置神經(jīng)網(wǎng)絡(luò)Simulink模塊初始條件getsiminit-GetneuralnetworkSimulinkblockinitialconditions獲得神經(jīng)網(wǎng)絡(luò)Simulink模塊初始條件neural-NeuralnetworkSimulinkblockset.神經(jīng)網(wǎng)絡(luò)Simulink模塊集Trainingfunctions.訓(xùn)練函數(shù)trainb-Batchtrainingwithweight&biaslearningrules.批處理具有權(quán)重和偏差學(xué)習(xí)規(guī)則的訓(xùn)練trainbfg-BFGSquasi-Newtonbackpropagation.BFGS擬牛頓反向傳播trainbr-BayesianRegulationbackpropagation.貝葉斯規(guī)則的反向傳播trainbu-Unsupervisedbatchtrainingwithweight&biaslearningrules.無(wú)監(jiān)管的批處理具有權(quán)重和偏差學(xué)習(xí)規(guī)則的訓(xùn)練trainbuwb-Unsupervisedbatchtrainingwithweight&biaslearningrules.無(wú)監(jiān)管的批處理具有權(quán)重和偏差學(xué)習(xí)規(guī)則的訓(xùn)練trainc-Cyclicalorderweight/biastraining.循環(huán)順序權(quán)重和偏差訓(xùn)練traincgb-ConjugategradientbackpropagationwithPowell-Bealerestarts.具有Powell-Beale重新開(kāi)始的共軛梯度反向傳播traincgf-ConjugategradientbackpropagationwithFletcher-Reevesupdates.具有Fletcher-Reeves更新的共軛梯度反向傳播traincgp-ConjugategradientbackpropagationwithPolak-Ribiereupdates.具有Polak-Ribiere更新的共軛梯度Polak-Ribieretraingd-Gradientdescentbackpropagation.梯度下降反向傳播traingda-Gradientdescentwithadaptivelrbackpropagation.具有自適應(yīng)LR的反向傳播梯度下降traingdm-Gradientdescentwithmomentum.具有動(dòng)量的梯度下降traingdx-Gradientdescentw/momentum&adaptivelrbackpropagation.梯度下降瓦特/動(dòng)量與自適應(yīng)LR的反向傳播trainlm-Levenberg-Marquardtbackpropagation.Levenberg-Marquardt反向傳播trainoss-Onestepsecantbackpropagation.一步割線(xiàn)反向傳播trainr-Randomorderweight/biastraining.隨機(jī)權(quán)重/偏差訓(xùn)練trainrp-RPROPbackpropagation.RPROP反向傳播trainru-Unsupervisedrandomorderweight/biastraining.無(wú)監(jiān)管隨機(jī)權(quán)重/偏差訓(xùn)練trains-Sequentialorderweight/biastraining.順序權(quán)重/偏差訓(xùn)練trainscg-Scaledconjugategradientbackpropagation.規(guī)?；曹椞荻确聪騻鞑lottingfunctions.繪圖函數(shù)plotconfusion-Plotclassificationconfusionmatrix.圖分類(lèi)混淆矩陣ploterrcorr-Plotautocorrelationoferrortimeseries.誤差自相關(guān)時(shí)間序列圖ploterrhist-Ploterrorhistogram.誤差直方圖plotfit-Plotfunctionfit.繪圖功能（函數(shù)）配合plotinerrcorr-Plotinputtoerrortimeseriescross-correlation.繪制輸入誤差時(shí)間序列互相關(guān)plotperform-Plotnetworkperformance.繪制網(wǎng)絡(luò)性能plotregression-Plotlinearregression.繪制線(xiàn)性回歸plotresponse-Plotdynamicnetworktime-seriesresponse.繪制動(dòng)態(tài)網(wǎng)絡(luò)時(shí)間序列反應(yīng)plotroc-Plotreceiveroperatingcharacteristic.繪制接收器操作特性plotsomhits-Plotself-organizingmapsamplehits.繪制自組織映射樣本采樣數(shù)plotsomnc-PlotSelf-organizingmapneighborconnections.繪制自組織映射鄰居關(guān)系plotsomnd-PlotSelf-organizingmapneighbordistances.繪制自組織映射鄰居距離plotsomplanes-Plotself-organizingmapweightplanes.繪制自組織映射權(quán)重位面plotsompos-Plotself-organizingmapweightpositions.繪制自組織映射權(quán)重位置plotsomtop-Plotself-organizingmaptopology.繪制自組織映射拓?fù)浣Y(jié)構(gòu)plottrainstate-Plottrainingstatevalues.繪制訓(xùn)練狀態(tài)值plotwb-PlotHintondiagramsofweightandbiasvalues.繪制權(quán)重和偏差值得Hinton圖Listsofotherneuralnetworkimplementationfunctions.神經(jīng)網(wǎng)絡(luò)其他重要函數(shù)的列表nnadapt-Adaptfunctions.適應(yīng)函數(shù)adaptwb-Sequentialorderincrementaltrainingw/learningfunctions.nnderivative-Derivativefunctions.派生（衍生）函數(shù)NeuralNetworkToolboxCalculationFunctions.神經(jīng)網(wǎng)絡(luò)工具箱計(jì)算函數(shù)Thecontentsofthisdirectorymaychangewithoutnotice.Thesefilesareimplementationfilesnotintendedfordirectuse.這個(gè)目錄的內(nèi)容可以改變故沒(méi)有公告，這些文件是安裝啟用文件并非直接可應(yīng)用的。nndistance-Distancefunctions.距離函數(shù)boxdist-Boxdistancefunction.Box距離函數(shù)dist-Euclideandistanceweightfunction.歐幾里得距離權(quán)值函數(shù)linkdist-Linkdistancefunction.Link距離函數(shù)mandist-Manhattandistancefunction.Manhattan距離函數(shù)nndivision-Divisionfunctions.除法函數(shù)divideblock-Partitionindicesintothreesetsusingblocksofindices.劃分指標(biāo)成為三個(gè)可用指標(biāo)塊集divideind-Partitionindicesintothreesetsusingspecifiedindices.劃分指標(biāo)成為三個(gè)可用指定指標(biāo)集divideint-Partitionindicesintothreesetsusinginterleavedindices.劃分指標(biāo)成為三個(gè)可用交叉存儲(chǔ)指標(biāo)集dividerand-Partitionindicesintothreesetsusingrandomindices.劃分指標(biāo)成為三個(gè)可用隨機(jī)指標(biāo)集dividetrain-Partitionindicesintotrainingsetonly.劃分指標(biāo)僅成為訓(xùn)練集nninitlayer-Initializelayerfunctions.初始化層函數(shù)initnw-Nguyen-Widrowlayerinitializationfunction.Nguyen-Widrow層初始化函initwb-By-weight-and-biaslayerinitializationfunction.附近（次要的）-權(quán)值和閾值層初始化函數(shù)nninitnetwork-Initializenetworkfunctions.初始化網(wǎng)絡(luò)函數(shù)initlay-Layer-by-layernetworkinitializationfunction.疊層網(wǎng)絡(luò)初始化函數(shù)nninitweight-Initializeweightfunctions.初始化權(quán)值函數(shù)Biasonlyinitializationfunctions僅閾值初始化函數(shù)initcon-Consciencebiasinitializationfunction.Conscience閾值初始化函數(shù)nnlearn-Learningfunctions.學(xué)習(xí)函數(shù)learncon-Consciencebiaslearningfunction.Conscience閾值學(xué)習(xí)函數(shù)learngd-Gradientdescentweight/biaslearningfunction.梯度下降權(quán)值/閾值學(xué)習(xí)函數(shù)learngdm-Gradientdescentw/momentumweight/biaslearningfunction.梯度下降w/動(dòng)量權(quán)值/閾值學(xué)習(xí)函數(shù)learnh-Hebbweightlearningrule.Hebb權(quán)值學(xué)習(xí)規(guī)則learnhd-Hebbwithdecayweightlearningrule.有衰減Hebb權(quán)值學(xué)習(xí)規(guī)則learnis-Instarweightlearningfunction.Instar權(quán)值學(xué)習(xí)函數(shù)learnk-Kohonenweightlearningfunction.Kohonen權(quán)值學(xué)習(xí)函數(shù)learnlv1-LVQ1weightlearningfunction.LVQ1權(quán)值學(xué)習(xí)函數(shù)learnlv2-LVQ2.1weightlearningfunction.LVQ2.1權(quán)值學(xué)習(xí)函數(shù)learnos-Outstarweightlearningfunction.Outstar權(quán)值學(xué)習(xí)函數(shù)learnp-Perceptronweight/biaslearningfunction.感知器的權(quán)值/偏差學(xué)習(xí)函數(shù)learnpn-Normalizedperceptronweight/biaslearningfunction.標(biāo)準(zhǔn)感知器的學(xué)習(xí)函數(shù)learnsom-Self-organizingmapweightlearningfunction.自組織映射權(quán)值學(xué)習(xí)函數(shù)learnsomb-LEARNSOMBatchself-organizingmapweightlearningfunction.LEARNSOM批處理自組織映射權(quán)值學(xué)習(xí)函數(shù)learnwh-Widrow-Hoffweight/biaslearningfunction.Widrow-Hoff權(quán)值/偏差學(xué)習(xí)函數(shù)nnnetinput-Netinputfunctions.網(wǎng)絡(luò)輸入函數(shù)netprod-Productnetinputfunction.Product網(wǎng)絡(luò)輸入函數(shù)netsum-Sumnetinputfunction.Sum網(wǎng)絡(luò)輸入函數(shù)nnperformance-Performancefunctions.性能函數(shù)mae-Meanabsoluteerrorperformancefunction.平均絕對(duì)誤差性能函數(shù)mse-Meansquarederrorperformancefunction.均方誤差性能函數(shù)sae-Sumabsoluteerrorperformancefunction.Sum絕對(duì)誤差性能函數(shù)sse-Sumsquarederrorperformancefunction.Sum均方誤差性能函數(shù)nnprocess-Processingfunctions.處理函數(shù)GeneralDataPreprocessing一般數(shù)據(jù)預(yù)處理fixunknowns-Processesmatrixrowswithunknownvalues.用未知值處理矩陣行mapminmax-Mapmatrixrowminimumandmaximumvaluesto[-11].映射矩陣行最小和最大值成[-1,1]mapstd-Mapmatrixrowmeansanddeviationstostandardvalues.映射矩陣行均值和偏差成標(biāo)準(zhǔn)值processpca-Processesrowsofmatrixwithprincipalcomponentanalysis.用主成分分析處理矩陣行removeconstantrows-Removematrixrowswithconstantvalues.用常量移去矩陣行removerows-Removematrixrowswithspecifiedindices.用指定指標(biāo)移去矩陣行DataPreprocessingforSpecificAlgorithms特定算法數(shù)據(jù)預(yù)處理lvqoutputs-DefinesettingsforLVQoutputs,withoutchangingvalues.為L(zhǎng)VQ定義設(shè)置，沒(méi)有值的改變nnsearch-Linesearchfunctions.線(xiàn)性搜索函數(shù)srchbac-One-dimensionalminimizationusingbacktracking.用回溯法一維最小化srchbre-One-dimensionalintervallocationusingBrent'smethod.用Brent法一維區(qū)間位置srchcha-One-dimensionalminimizationusingthemethodofCharalambous.用Charalambous方法一維最小化srchgol-One-dimensionalminimizationusinggoldensectionsearch.用黃金分割收索一維最小化srchhyb-One-dimensionalminimizationusingahybridbisection-cubicsearch.用平分三次收索一維最小化nntopology-Topologyfunctions.拓?fù)浜瘮?shù)gridtop-Gridlayertopologyfunction.網(wǎng)格層拓?fù)浜瘮?shù)hextop-Hexagonallayertopologyfunction.六角形層拓?fù)浜瘮?shù)randtop-Randomlayertopologyfunction.隨機(jī)層拓?fù)浜瘮?shù)tritop-Trianglelayertopologyfunction.三角形層拓?fù)浜瘮?shù)nntransfer-Transferfunctions.傳遞函數(shù)compet-Competitivetransferfunction.競(jìng)爭(zhēng)傳遞函數(shù)elliotsig-Elliotsigmoidtransferfunction.ElliotS型傳遞函數(shù)hardlim-Positivehardlimittransferfunction.正的硬界限傳遞函數(shù)hardlims-Symmetrichardlimittransferfunction.對(duì)稱(chēng)硬界