回归预测 - MATLAB实现GRU(门控循环单元)多输入单输出（完整源码和数据）

%% clc;clear; warning off; %% 导入数据 %% data = xlsread('mata.xlsx', 'Sheet1', 'A3:M1250'); % 输入数据 input =data(:,1:12)'; output=data(:,13)'; nwhole =size(data,1); train_ratio=0.9; ntrain=round(nwhole*train_ratio); ntest =nwhole-ntrain; % 准备输入和输出训练数据 input_train =input(:,1:ntrain); output_train=output(:,1:ntrain); % 准备测试数据 input_test =input(:, ntrain+1:ntrain+ntest); output_test=output(:,ntrain+1:ntrain+ntest); %% 归一化（全部特征 均归一化） [inputn_train,inputps] =mapminmax(input_train); [outputn_train,outputps]=mapminmax(output_train); inputn_test =mapminmax('apply',input_test,inputps); outputn_test=mapminmax('apply',output_test,outputps); %% GRU 层设置，参数设置 inputSize = size(inputn_train,1); %数据输入x的特征维度 outputSize = size(outputn_train,1); %数据输出y的维度 numhidden_units1=60; numhidden_units2=180; numhidden_units3=60; %% gru layers = [ ... sequenceInputLayer(inputSize) %输入层设置 gruLayer(numhidden_units1,'name','hidden1') %学习层设置(cell层） dropoutLayer(0.2,'name','dropout_1') gruLayer(numhidden_units2,'Outputmode','sequence','name','hidden2') dropoutLayer(0.3,'name','dropout_2') gruLayer(numhidden_units3,'name','hidden3') dropoutLayer(0.2,'name','dropout_3') fullyConnectedLayer(outputSize) % 全连接层设置（影响输出维度） regressionLayer('name','out')]; %% trainoption(gru) opts = trainingOptions('adam', ... 'MaxEpochs',200, ... 'GradientThreshold',1,... 'ExecutionEnvironment','cpu',... 'InitialLearnRate',0.005, ... 'LearnRateSchedule','piecewise', ... 'LearnRateDropPeriod',100, ... % epoch后学习率更新 'LearnRateDropFactor',0.8, ... 'Verbose',0, ... 'Plots','training-progress'... ); %% GRU网络训练 tic GRUnet = trainNetwork(inputn_train ,outputn_train ,layers,opts); toc; [GRUnet,GRUoutputr_train]= predictAndUpdateState(GRUnet,inputn_train); GRUoutput_train = mapminmax('reverse',GRUoutputr_train,outputps); %% GRU测试数据 %% %网络测试输出 [GRUnet,GRUoutputr_test] = predictAndUpdateState(GRUnet,inputn_test); %网络输出反归一化 GRUoutput_test= mapminmax('reverse',GRUoutputr_test,outputps); %% GRU数据输出 %% %------------------------------------------------------------------------------------- error_test=GRUoutput_test'-output_test'; pererror_test=error_test./output_test'; error=error_test'; pererror=pererror_test'; avererror=sum(abs(error))/(ntest); averpererror=sum(abs(pererror))/(ntest); RMSE = sqrt(mean((error).^2)); disp('GRU网络预测绝对平均误差MAE'); disp(avererror); disp('GRU网络预测平均绝对误差百分比MAPE'); disp(averpererror) disp('GRU网络预测均方根误差RMSE') disp(RMSE) %% GRU数据可视化分析 %测试数据 figure() plot(GRUoutput_test,'m-.','linewidth',1.5) hold on plot(output_test,'k--','linewidth',1) legend( '预测测试数据','实际分析数据','Location','NorthWest','FontName','宋体'); title('GRU网络模型结果及真实值','fontsize',12,'FontName','宋体') xlabel('时间','fontsize',12,'FontName','宋体'); ylabel('数值','fontsize',12,'FontName','宋体'); %------------------------------------------------------------------------------------- figure() stairs(pererror_test,'-.','Color',[255 50 0]./255,'linewidth',1.5) legend('GRU网络测试相对误差','Location','NorthEast','FontName','宋体') title('GRU网络预测相对误差','fontsize',12,'FontName','宋体') ylabel('误差','fontsize',12,'FontName','宋体') xlabel('样本','fontsize',12,'FontName','宋体') %-------------------------------------------------------------------------------------