基于OLS算法的预测1.doc

程序(图在下面): %基于OLS 的RBF 网设计算法 %%------------------------------ clc; clear all; clc; load data.mat;%约简后的总体样本 ceshiyangben=[data(1:2:238,1),data(1:2:238,2),data(1:2:238,3),data(1:2:238,4),data(1:2:238,5),data(1:2:238,6),data(1:2:238,7),data(1:2:238,8)]; ceshiyangbenjieguo=[data(1:2:238,9)]; yuceyangben=[data(2:2:119,1),data(2:2:119,2),data(2:2:119,3),data(2:2:119,4),data(2:2:119,5),data(2:2:119,6),data(2:2:119,7),data(2:2:119,8)]; yuceyangbenjieguo=[data(2:2:119,9)]; SamNum = 119; % 训练样本数 TestSamNum = 59; % 测试样本数 SP = 0.6; % 隐节点扩展常数 ErrorLimit = 0.01; % 目标误差 % 根据目标函数获得样本输入输出 SamIn = ceshiyangben; SamOut = ceshiyangbenjieguo; TestSamIn = yuceyangben; TestSamOut = yuceyangbenjieguo; [InDim,MaxUnitNum] = size(SamIn); % 样本输入维数和最大允许隐节点数 % 计算隐节点输出阵 Distance = dist(SamIn,SamIn); HiddenUnitOut = radbas(Distance/SP); PosSelected = []; VectorsSelected = []; HiddenUnitOutSelected = []; ErrHistory = []; % 用于记录每次增加隐节点后的训练误差 VectorsSelectFrom = HiddenUnitOut; dd = sum((SamOut.*SamOut)); % 计算各隐节点输出矢量与目标输出矢量的夹角平方值 for k = 1 : MaxUnitNum PP = sum(VectorsSelectFrom.*VectorsSelectFrom); Denominator = dd * PP; [xxx,SelectedNum] = size(PosSelected); if SelectedNum0 [lin,xxx] = size(Denominator); Denominator(:,PosSelected) = ones(lin,1); end Angle = ((SamOut*VectorsSelectFrom) .^ 2) ./ Denominator; % 选择具有最大投影的矢量，得到相应的数据中心 [value,pos] = max(Angle); PosSelected = [PosSelected pos]; % 计算RBF网训练误差 HiddenUnitOutSelected = [HiddenUnitOutSelected;HiddenUnitOut(pos,:)]; HiddenUnitOutEx = [HiddenUnitOutSelected;ones(1,SamNum)]; W2Ex = SamOut*pinv(HiddenUnitOutEx); % 用广义逆求广义输出权值 W2 = W2Ex(:,1:k); % 得到输出权值 B2 = W2Ex(:,k+1); % 得到偏移 NNOut = W2*HiddenUnitOutSelected+B2; % 计算RBF网输出 SSE = sumsqr(SamOut - NNOut); % 记录每次增加隐节点后的训练误差 ErrHistory = [ErrHistory SSE]; if SSE ErrorLimit break, end