模糊神經(jīng)網(wǎng)絡(luò)算法之MATLAB實(shí)現(xiàn)

1、窗體頂端用戶名:   密碼: 登錄 窗體底端注冊 礦業(yè)人生的空間 主頁博客相冊|個人檔案 |好友 |i貼吧  查看文章 【轉(zhuǎn)】 模糊神經(jīng)網(wǎng)絡(luò)算法之MATLAB實(shí)現(xiàn)2011-05-07 12:59轉(zhuǎn)載自 分享最終編輯 whutzxlfunction retstr = FnnTrain(dt,ld,tt,sp) retstr=-1; % 輸入?yún)?shù)賦值開始 % % 方便調(diào)試程序用，程序調(diào)試時去掉這部分的注釋 % dt=4;      %學(xué)習(xí)閾值 % ld=0.05;   

2、 %學(xué)習(xí)進(jìn)度 % tt=10;      %訓(xùn)練次數(shù) % sp='datasample.txt'    %一個點(diǎn)的監(jiān)測數(shù)據(jù) % 輸入?yún)?shù)賦值結(jié)束 % global recordDimention; global sampleNumber; global weightNumber; global distanceThread; global WW; global learningDegree; global epochsNumber; distanceThread=dt; learningDegree

3、=ld; traintimes=tt; A=load(sp);    Arow Acol=size(A); %樣本個數(shù) sampleNumber=Arow; recordDimention=Acol; disp(sampleNumber); WW=A(1,:); WW=WW 1; weightNumber=1; epochsNumber=1;    for jj=2:1:sampleNumber       TrainNN2(A(jj,:); end for jt=1:traintimes-1  

4、;     for jt2=1:sampleNumber           TrainNN2(A(jj,:);       end end % 將訓(xùn)練結(jié)果寫入權(quán)值文件 dlmwrite('dataw.dat',WW,'t'); % % 訓(xùn)練子函數(shù) % function TrainNN2(a) global recordDimention; global sampleNumber; g

5、lobal weightNumber; global distanceThread; global WW; global learningDegree; global epochsNumber; Ldistance=zeros(2,weightNumber); for j1=1:weightNumber       Ldistance(2,j1)=j1; end % %輸入輸出空間的模糊分割 for j2=1:weightNumber       Lx=0;    

6、   for j3=1:recordDimention           Lx=Lx + (a(j3) - WW(j2,j3) .* (a(j3) - WW(j2,j3);       end       Ldistance(1,j2)=sqrt(Lx); end % %計算模糊空間的距離   Lx1=Ldistance(1,1); Lx2=Ldistance(2,

7、1); for j4=2:weightNumber       if(Ldistance(1,j4)<Lx1)           Lx1=Ldistance(1,j4);           Lx2=Ldistance(2,j4);       end end % %修正模糊規(guī)則 updateW=0; nowW

8、eight=Lx2; if(Lx1<=distanceThread)       for j6=1:recordDimention           WW(nowWeight,j6)=WW(nowWeight,j6) + learningDegree .* (a(j6) - WW(nowWeight,j6);       end       WW(n

9、owWeight,recordDimention+1)=WW(nowWeight,recordDimention+1)+1;       updateW=1; end      if( updateW=0)       weightNumber=weightNumber+1;       a=a 1;       WW=WW;a; end類別：默認(rèn)分類 |

10、| 添加到搜藏 | 分享到i貼吧 | 瀏覽(63) | 評論 (0)  上一篇：考拉BP神經(jīng)網(wǎng)絡(luò)的matlab實(shí)現(xiàn)學(xué)習(xí).    下一篇：引入簡單變異算子的粒子群算法尋. 相關(guān)文章：Kohonen算法實(shí)現(xiàn)自組織特征映射.BP神經(jīng)網(wǎng)絡(luò)算法matlab實(shí)現(xiàn)BP神經(jīng)網(wǎng)絡(luò)的設(shè)計實(shí)例(MATLAB編.關(guān)于神經(jīng)網(wǎng)絡(luò)(matlab)歸一化的整.Matlab中BP神經(jīng)網(wǎng)絡(luò)的構(gòu)建和初始.Matlab的神經(jīng)網(wǎng)絡(luò)工具箱入門BP神經(jīng)網(wǎng)絡(luò)matlab實(shí)現(xiàn)的基本步驟Matlab的神經(jīng)網(wǎng)絡(luò)工具箱實(shí)用指南.提供一個Matlab的BP神經(jīng)網(wǎng)絡(luò)的基.Matlab的神經(jīng)網(wǎng)絡(luò)工具箱實(shí)用指南.更多>> 最近讀者：登錄后，您就出現(xiàn)在這里。  yusiking52xibeifeiying