SCI 中科院1区算法应用：雪消融优化SAO-CNN-BiGRU-Attention融合注意力机制预测程序代码！直接运行！

%% 导入特征数据、当天的风速数据data = xlsread('特征序列及实际值.xlsx');Features   = data(1:18,:);                             %% 特征输入  ：75天，每天24小时，每小时一个采样点，共计75*24=1800小时，18个特征数据Wind_data  = data(19,:);                               %% 实际值输出：75天，每天24小时，每小时一个采样点，共计75*24=1800小时的风速数据
%%  数据平铺为4-DLP_Features =  double(reshape(Features,18,24,1,75));   %% 特征数据格式为18*24*1*75，分别对应18特征24小时，75天LP_WindData  = double(reshape(Wind_data,24,1,1,75));   %% 实际数据格式为24*1*1*75 ，分别对应24小时，75天
%% 格式转换为cellNumDays  = 75;                                         %% 数据总天数为 75天for i=1:NumDays    FeaturesData{1,i} = LP_Features(:,:,1,i);end
for i=1:NumDays    RealData{1,i} = LP_WindData(:,:,1,i);end
%% 划分数据XTrain = FeaturesData(:,1:73);                         %% 训练集输入为 1-73   天的特征YTrain = RealData(:,2:74);                             %% 训练集输出为 2-74天 的实际值                
XTest  = cell2mat(FeaturesData(: , 74));               %% 测试集输入第  74    天的特征Ytest  = cell2mat(RealData(: , 75));                   %% 测试集输出为第 75天 的实际值
%% 将优化目标参数传进来的值 转换为需要的超参数learning_rate = x(1);            %% 学习率KerlSize = round(x(2));          %% 卷积核大小NumNeurons = round(x(3));        %% 神经元个数
%% 网络搭建lgraph = layerGraph();
% 添加层分支% 将网络分支添加到层次图中。每个分支均为一个线性层组。tempLayers = sequenceInputLayer([18 24 1],"Name","sequence");lgraph = addLayers(lgraph,tempLayers);
tempLayers = [    convolution2dLayer(KerlSize,32,"Name","conv","Padding","same")    batchNormalizationLayer("Name","batchnorm")    reluLayer("Name","relu")    maxPooling2dLayer([3 3],"Name","maxpool","Padding","same")    flattenLayer("Name","flatten_1")    fullyConnectedLayer(25,"Name","fc_1")];lgraph = addLayers(lgraph,tempLayers);
tempLayers = flattenLayer("Name","flatten");lgraph = addLayers(lgraph,tempLayers);
tempLayers = gruLayer(NumNeurons,"Name","gru1");lgraph = addLayers(lgraph,tempLayers);
tempLayers = [    FlipLayer("flip3")    gruLayer(NumNeurons,"Name","gru2")];lgraph = addLayers(lgraph,tempLayers);
tempLayers = [    concatenationLayer(1,3,"Name","concat")    selfAttentionLayer(1,50,"Name","selfattention")    fullyConnectedLayer(24,"Name","fc")    regressionLayer("Name","regressionoutput")];lgraph = addLayers(lgraph,tempLayers);
% 清理辅助变量clear tempLayers;
% 连接层分支% 连接网络的所有分支以创建网络图。lgraph = connectLayers(lgraph,"sequence","conv");lgraph = connectLayers(lgraph,"sequence","flatten");lgraph = connectLayers(lgraph,"flatten","gru1");lgraph = connectLayers(lgraph,"flatten","flip3");lgraph = connectLayers(lgraph,"gru1","concat/in1");lgraph = connectLayers(lgraph,"gru2","concat/in2");lgraph = connectLayers(lgraph,"fc_1","concat/in3");
%% 设置训练参数options = trainingOptions('adam', ...   % adam 梯度下降算法    'MaxEpochs',300, ...                % 最大训练次数 300    'GradientThreshold',1,...           % 渐变的正阈值 1    'ExecutionEnvironment','cpu',...    % 网络的执行环境 cpu    'InitialLearnRate',learning_rate,...% 初始学习率 0.01    'LearnRateSchedule','none',...      % 训练期间降低整体学习率的方法 不降低    'Shuffle','every-epoch',...         % 每次训练打乱数据集    'SequenceLength',24,...             % 序列长度 24    'MiniBatchSize',20,...              % 训练批次大小 每次训练样本个数20    'Verbose',true);                    % 有关训练进度的信息不打印到命令窗口中
%% 训练网络[net,info] = trainNetwork(XTrain,YTrain, lgraph, options);