% 讀入訓練數據和測試數據 
Input = []; 
Output = []; 
str = {'Test','Check'}; 
Data = textread([str{1},'.txt']); 
% 讀訓練數據 
Input = Data(:,1:end-1); 
% 取數據表的前五列（主從成分） 
Output = Data(:,end); 
% 取數據表的最后一列（輸出值） 
Data = textread([str{2},'.txt']); 
% 讀測試數據 
CheckIn = Data(:,1:end-1); 
% 取數據表的前五列（主從成分） 
CheckOut = Data(:,end); 
% 取數據表的最后一列（輸出值） 
Input = Input'; 
Output = Output'; 
CheckIn = CheckIn'; 
CheckOut = CheckOut'; 
% 矩陣轉置 
[Input,minp,maxp,Output,mint,maxt] = premnmx(Input,Output); 
% 標準化數據 

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
% 神經網絡參數設置 
%====可以修正處 
Para.Goal = 0.0001; 
% 網絡訓練目標誤差 
Para.Epochs = 800; 
% 網絡訓練次數 
Para.LearnRate = 0.1; 
% 網絡學習速率 
%==== 
Para.Show = 5; 
% 網絡訓練顯示間隔 
Para.InRange = repmat([-1 1],size(Input,1),1); 
% 網絡的輸入變量區間 
Para.Neurons = [size(Input,1)*2+1 1]; 
% 網絡后兩層神經元配置 
Para.TransferFcn= {'logsig' 'purelin'}; 
% 各層的閾值函數 
Para.TrainFcn = 'trainlm'; 
% 網絡訓練函數賦值 
% traingd : 梯度下降后向傳播法 
% traingda : 自適應學習速率的梯度下降法 
% traingdm : 帶動量的梯度下降法 
% traingdx : 
% 帶動量，自適應學習速率的梯度下降法 
Para.LearnFcn = 'learngdm'; 
% 網絡學習函數 
Para.PerformFcn = 'sse'; 
% 網絡的誤差函數 
Para.InNum = size(Input,1); 
% 輸入量維數 
Para.IWNum = Para.InNum*Para.Neurons(1); 
% 輸入權重個數 
Para.LWNum = prod(Para.Neurons); 
% 層權重個數 
Para.BiasNum = sum(Para.Neurons); 
% 偏置個數 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
Net = newff(Para.InRange,Para.Neurons,Para.TransferFcn,... 
Para.TrainFcn,Para.LearnFcn,Para.PerformFcn); 
% 建立網絡 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
Net.trainParam.show = Para.Show; 
% 訓練顯示間隔賦值 
Net.trainParam.goal = Para.Goal; 
% 訓練目標誤差賦值 
Net.trainParam.lr = Para.LearnRate; 
% 網絡學習速率賦值 
Net.trainParam.epochs = Para.Epochs; 
% 訓練次數賦值 
Net.trainParam.lr = Para.LearnRate; 

Net.performFcn = Para.PerformFcn; 
% 誤差函數賦值 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
% 調試 
Out1 =sim(Net,Input); 
% 仿真剛建立的網絡 
Sse1 =sse(Output-Out1); 
% 剛建立的網絡誤差 

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
[Net TR] = train(Net,Input,Output); 
% 訓練網絡并返回 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 
Out3 =sim(Net,Input); 
% 對學習訓練后的網絡仿真