MATLAB: 基于差分进化算法（DE）的神经网络优化UCI数据集

%% %%%%%%%%%%%%基于DE优化神经网络%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %需求：优化参数、网络结构 % 1. 优化参数权值、阈值 % 2. 优化隐含层数 %% 进化算法优化神经网络，其中最主要面临的问题在于误差函数的选取作为适应度函数的问题 clc clear close all %% %%%%%%%%%%%%%%%%%%%%%%%确定网络各层神经元个数%%%%%%%%%%%%%%%%%%%%%%%%%%% %所选的数据应该进行特征筛选，选择最有效的数据进行处理 % load('UCI测试数据集/iris.mat') % data1=iris(:,1:end-1); % load ('UCI测试数据集/Wine.mat') % data1=Wine(:,2:end-1); % load('UCI测试数据集/Balancescale.mat') % data1=Balancescale(:,2:end); % load('UCI测试数据集/breast.mat') % data1=breast(:,1:end-1); % load('UCI测试数据集/diabetes.mat') % data1=diabetes(:,1:end-1); % load('UCI测试数据集/haberman.mat') % data1=haberman(:,1:end-1); % load('UCI测试数据集/heart.mat') % data1=heart(:,1:end-1); % load('UCI测试数据集/ionosphere.mat') % data1=ionosphere(:,1:end-1); load('UCI测试数据集/Seeds.mat') data1=Seeds(:,2:end); % load('UCI测试数据集/zoo.mat') % data1=zoo(:,1:end-1); [row,col]=size(data1); global insize hidesize outsize insize = col; %输入层神经元数目--------根据数据集所决定的 hidesize = 4; %隐含层神经元数目--------由自己设定 outsize = 1; %输出层神经元数目--------输入神经元与输出神经元的个数相同 %% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%导入特征数据数据%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% input=data1'; testdata=input(:,120:140); input_traindata=input(:,1:row)'; %% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%导入标签%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % output_train=iris(1:row,end)'; % output_test=iris(40:70,end)'; % output_train=Wine(1:row,1)'; % output_test=Wine(120:150,1)'; % output_train=Balancescale(1:row,1)'; % output_test=Balancescale(120:140,1)'; % output_train=breast(1:row,end)'; % output_test=breast(120:140,end)'; % output_train=diabetes(1:row,end)'; % output_test=diabetes(300:320,end)'; % output_train=haberman(1:row,end)'; % output_test=haberman(200:220,end)'; % output_train=heart(1:row,end)'; % output_test=heart(120:140,end)'; % output_train=ionosphere(1:row,end)'; % output_test=ionosphere(120:140,end)'; output_train=Seeds(1:row,1)'; output_test=Seeds(120:140,1)'; % output_train=zoo(1:row,end)'; % output_test=zoo(80:100,end)'; [output_testLable,outputps]=mapminmax(output_train); %训练样本输出值归一化到[-1,1]之间 %% %%%%%%%%使用HSDE进行优化参数----训练模型%%%%%%%%%%%%%%%%%%% output_traindata=output_testLable'; %归一化后的结果 popsize=40; %种群大小 invidualsize=hidesize*(insize +outsize+1)+outsize; iter=1000; %迭代次数 [best,optvalue]=DE(popsize,invidualsize,iter,input_traindata,output_traindata); loglog(optvalue,'m-','linewidth',1.5); xlabel('进化代数');ylabel('样本总准则函数值') %% %%%%%%%%%%%%设计将优化出来的值进行转化为权值和阈值，存档%%%%%%%%%%%%%%%%%%%%%% c=best(1:hidesize*(insize +outsize)); p=zeros(hidesize,insize +outsize); pp=zeros(outsize,insize +outsize); for i=1:size(c,2)/(insize+1) for j=1:(insize+1) p(i,j)=c(1); c(1)=[]; end end W1=p(:,1:insize); B1=p(:,(insize+1)); d=best(hidesize*(insize +outsize)+1:end); for i=1:size(d,2)/(hidesize+1) for j=1:hidesize+1 pp(i,j)=d(1); d(1)=[]; end end W2=pp(:,1:hidesize); B2=pp(:,hidesize+1); %% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%测试模型%%%%%%%%%%%%%%%%%%%%%%%%% accuracy_numbers=0; tempyout = zeros(outsize,size(testdata,2)); for i = 1:size(testdata,2) testdata_x = testdata(:,i); tesetdata_hidein = W1*testdata_x+B1;%隐含层输入值 testdata_hideout = zeros(hidesize,1);%隐含层输出值 for j = 1:hidesize testdata_hideout(j) = SigmoidFun(tesetdata_hidein(j)); end testdata_yin= W2*testdata_hideout+B2;%输出层输入值 testdata_yout = zeros(outsize,1); for j = 1:outsize testdata_yout(j) = SigmoidFun( testdata_yin(j)); end tempyout(:,i) =testdata_yout; %最终测试集的输出值-----这里需要将其与测试集的类别进行对比 % if testdata_yout>0 && testdata_yout<=1.5 % result_out(i)=1; % elseif testdata_yout>1.5 && testdata_yout<=2.5 % if result_out(i)==testdata(3,i) % accuracy_numbers= accuracy_numbers+1; % end % hold on; % else % scatter(testdata_x(1),testdata_x(2),'g') %给分类做一个边缘分割线---怎样处理边缘框线 % result_out(i)=-1; % if result_out(i)==testdata(3,i) % accuracy_numbers= accuracy_numbers+1; % end % hold on; % title('BP神经网络测试模型分类效果图'); % end end test_simu=mapminmax('reverse', tempyout,outputps); %把仿真得到的数据还原为原始的数量级 error=abs(test_simu-output_test); %预测值和真实值的误差 %% %%%%%%%%%%判断函数的预测精度%%%%%%%%%%%%%%%%%%%%% successful_numbers=0; unsuccessful_numbers=0; successful_accuracy=abs(output_test-test_simu); for i=1:size(successful_accuracy,2) if successful_accuracy(:,i)/output_test(:,i)<0.05 successful_numbers= successful_numbers+1; else unsuccessful_numbers= unsuccessful_numbers+1; end end prec_accuracy= successful_numbers/size(output_test,2); disp(['预测精度为：',num2str( prec_accuracy)]); %% 第十步 真实值与预测值误差比较 figure(2) plot(output_test(1,:),'bo-') hold on plot(test_simu(1,:),'r*-') hold on plot(error(1,:),'square','MarkerFaceColor','b') legend('真实值','预测值','误差'); xlabel('测试集个数'); ylabel('输出值');