BiTCN-LSTM双向时间卷积长短期记忆神经网络时间序列预测（Matlab完整源码和数据）

% 清空环境变量 warning off % 关闭报警信息 close all % 关闭开启的图窗 clear % 清空变量 clc %% 导入数据 data = readmatrix('风电场预测.xlsx'); data = data(5665:8640,12); %选取3月份数据 [h1,l1]=data_process(data,24); %步长为24，采用前24个时刻的温度预测第25个时刻的温度 res = [h1,l1]; num_samples = size(res,1); %样本个数 % 训练集和测试集划分 outdim = 1; % 最后一列为输出 num_size = 0.7; % 训练集占数据集比例 num_train_s = round(num_size * num_samples); % 训练集样本个数 f_ = size(res, 2) - outdim; % 输入特征维度 P_train = res(1: num_train_s, 1: f_)'; T_train = res(1: num_train_s, f_ + 1: end)'; M = size(P_train, 2); P_test = res(num_train_s + 1: end, 1: f_)'; T_test = res(num_train_s + 1: end, f_ + 1: end)'; N = size(P_test, 2); % 数据归一化 [p_train, ps_input] = mapminmax(P_train, 0, 1); p_test = mapminmax('apply', P_test, ps_input); [t_train, ps_output] = mapminmax(T_train, 0, 1); t_test = mapminmax('apply', T_test, ps_output); % 格式转换 for i = 1 : M vp_train{i, 1} = p_train(:, i); vt_train{i, 1} = t_train(:, i); end for i = 1 : N vp_test{i, 1} = p_test(:, i); vt_test{i, 1} = t_test(:, i); end %% 创建网络 % 创建输入层 % 网络参数设置 filterSize = 3; % 滤波器大小 dropoutFactor = 0.1; numBlocks = 3; numFilters = 30; % 滤波器个数 NumNeurons = 25; % 神经元个数 layer = sequenceInputLayer(f_, Normalization = "rescale-symmetric", Name = "input"); % 创建网络图 lgraph = layerGraph(layer); outputName = layer.Name; % 建立网络结构 -- 残差块 for i = 1 : numBlocks % 膨胀因子 dilationFactor = 2^(i-1); % 创建TCN正向支路 layers = [ convolution1dLayer(filterSize, numFilters, DilationFactor = dilationFactor, Padding = "causal", Name="conv1_" + i) % 一维卷积层 layerNormalizationLayer % 层归一化 spatialDropoutLayer(dropoutFactor) % 空间丢弃层 convolution1dLayer(filterSize, numFilters, DilationFactor = dilationFactor, Padding = "causal") % 一维卷积层 layerNormalizationLayer % 层归一化 reluLayer % 激活层 spatialDropoutLayer(dropoutFactor) % 空间丢弃层 additionLayer(4, Name = "add_" + i) ]; % 添加残差块到网络 lgraph = addLayers(lgraph, layers); % 连接卷积层到残差块 lgraph = connectLayers(lgraph, outputName, "conv1_" + i); % 创建 TCN反向支路flip网络结构 Fliplayers = [ FlipLayer("flip_" + i) % 反向翻转 convolution1dLayer(1, numFilters, Name = "convSkip_"+i); % 反向残差连接 convolution1dLayer(filterSize, numFilters, DilationFactor = dilationFactor, Padding = "causal", Name="conv2_" + i) % 一维卷积层 layerNormalizationLayer % 层归一化 spatialDropoutLayer(dropoutFactor) % 空间丢弃层 convolution1dLayer(filterSize, numFilters, DilationFactor = dilationFactor, Padding = "causal") % 一维卷积层 layerNormalizationLayer % 层归一化 reluLayer % 激活层 spatialDropoutLayer(dropoutFactor, Name="drop" + i) % 空间丢弃层 ]; % 添加 flip 网络结构到网络 lgraph = addLayers(lgraph, Fliplayers); % 连接 flip 卷积层到残差块 lgraph = connectLayers(lgraph, outputName, "flip_" + i); lgraph = connectLayers(lgraph, "drop" + i, "add_" + i + "/in3"); lgraph = connectLayers(lgraph, "convSkip_"+i, "add_" + i + "/in4"); % 残差连接 -- 首层 if i == 1 % 建立残差卷积层 % Include convolution in first skip connection. layer = convolution1dLayer(1,numFilters,Name="convSkip"); lgraph = addLayers(lgraph,layer); lgraph = connectLayers(lgraph,outputName,"convSkip"); lgraph = connectLayers(lgraph,"convSkip","add_" + i + "/in2"); else lgraph = connectLayers(lgraph,outputName,"add_" + i + "/in2"); end % Update layer output name. outputName = "add_" + i; end tempLayers = flattenLayer("Name","flatten"); lgraph = addLayers(lgraph,tempLayers); tempLayers = [ lstmLayer(35,'Name','lstmlayer') fullyConnectedLayer(outdim,"Name","fc") regressionLayer("Name","regressionoutput")]; lgraph = addLayers(lgraph,tempLayers); lgraph = connectLayers(lgraph,outputName,"flatten"); lgraph = connectLayers(lgraph,"flatten","lstmlayer"); %% 训练参数设置 options = trainingOptions('adam', ... % 优化算法Adam 'MaxEpochs', 150, ... % 最大训练次数 'GradientThreshold', 1, ... % 梯度阈值 'InitialLearnRate', 0.01, ... % 初始学习率 'LearnRateSchedule', 'piecewise', ... % 学习率调整 'LearnRateDropPeriod', 120, ... % 训练后开始调整学习率 'LearnRateDropFactor',0.1, ... % 学习率调整因子 'L2Regularization', 0.001, ... % 正则化参数 'ExecutionEnvironment', 'cpu',... % 训练环境 'Verbose', 1, ... % 关闭优化过程 'Plots', 'none'); % 画出曲线 %% 训练模型 net = trainNetwork(vp_train,vt_train,lgraph,options); %% 查看网络结构 figure plot(lgraph) title("BiTCN-LSTM网络模型") set(gcf,'color','w') %analyzeNetwork(net);% 查看网络结构 % 预测 t_sim1 = predict(net, vp_train); t_sim2 = predict(net, vp_test); % 数据反归一化 T_sim1 = mapminmax('reverse', t_sim1, ps_output); T_sim2 = mapminmax('reverse', t_sim2, ps_output); T_train1 = T_train; T_test2 = T_test; % 数据格式转换 T_sim1 = cell2mat(T_sim1);% cell2mat将cell元胞数组转换为普通数组 T_sim2 = cell2mat(T_sim2); % 指标计算 disp('…………BiTCN-LSTM训练集误差指标…………') [mae1,rmse1,mape1,error1]=calc_error(T_train1,T_sim1'); fprintf('\n') figure('Position',[200,300,600,200]) plot(T_train1); hold on plot(T_sim1') legend('真实值','预测值') title('BiTCN-LSTM训练集预测效果对比') xlabel('样本点') ylabel('发电功率') disp('…………BiTCN-LSTM测试集误差指标…………') [mae2,rmse2,mape2,error2]=calc_error(T_test2,T_sim2'); fprintf('\n') figure('Position',[200,300,600,200]) plot(T_test2); hold on plot(T_sim2') legend('真实值','预测值') title('BiTCN-LSTM预测集预测效果对比') xlabel('样本点') ylabel('发电功率') figure('Position',[200,300,600,200]) plot(T_s