【TCN分类】基于双向时间卷积神经网络BiTCN实现故障诊断附matlab代码.rar

%% 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 % 清空命令行 %% 导入数据 res = xlsread('数据集.xlsx'); % rng(0); %% 数据分析 num_size = 0.7; % 训练集占数据集比例 outdim = 1; % 最后一列为输出 num_class = length(unique(res(:,end))); % 计算类别数 num_samples = size(res, 1); % 样本个数 kim = size(res, 2)-1; % 样本个数 res = res(randperm(num_samples), :); % 打乱数据集（不希望打乱时，注释该行） 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 flag_conusion = 1; % 打开混淆矩阵 %% 划分训练集和测试集 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 = categorical(T_train); t_test = categorical(T_test ); %% 设置网络参数 numFilters = 8; %卷积核个数 简单数据可以减少卷积和数量 filterSize = 2; %卷积核大小 dropoutFactor = 0.2; % 空间丢失因子 numBlocks = 2; %残差块个数 numFeatures = f_; %% 建立模型 layer = sequenceInputLayer(numFeatures,Normalization="rescale-symmetric",Name="input"); %% 加入空白网络 lgraph = layerGraph(layer); outputName = layer.Name; %% 残差网络 for i = 1:numBlocks 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"); if i == 1 layer = convolution1dLayer(1,numFilters,Name="convSkip"); %建立残差卷积 lgraph = addLayers(lgraph,layer); %残差卷积层 lgraph = connectLayers(lgraph,outputName,"convSkip"); %残差卷积层 lgraph = connectLayers(lgraph,"convSkip","add_" + i + "/in2"); %跳跃残差层 else % lgraph = connectLayers(lgraph, "drop" + i, "add_" + i + "/in2"); lgraph = connectLayers(lgraph,outputName,"add_" + i + "/in2"); end %% 更新输出层名字 outputName = "add_" + i; end %% 网络输出层 layers = [ fullyConnectedLayer(num_class,'Name',"fc") softmaxLayer classificationLayer]; lgraph = addLayers(lgraph,layers); lgraph = connectLayers(lgraph,outputName,"fc"); %% 参数设置 options = trainingOptions('adam', ... % Adam 梯度下降算法 'MaxEpochs', 500,... % 最大训练次数 'MiniBatchSize', 32, ... % 批大小 'InitialLearnRate', 0.01,... % 初始学习率为0.01 'L2Regularization', 0.01,... % L2正则化参数 'ExecutionEnvironment','cpu',... % 网络的执行环境 cpu 'Shuffle', 'every-epoch',... % 每次训练打乱数据集 'ValidationPatience', Inf,... % 关闭验证 'Plots', 'training-progress',... % 画出曲线 可关闭 'Verbose', false); tic %% 训练 [net , info]= trainNetwork(p_train, t_train, lgraph, options); run_time = toc; %% 输出训练时间 disp(['TCN的模型训练时间为： ', num2str(run_time),'s']) %% 查看网路 analyzeNetwork(lgraph); %% 绘制训练loss损失图 figure plot(1: length(info.TrainingAccuracy), info.TrainingAccuracy, 'b-', 'LineWidth', 1) legend('TCN训练准确率') xlabel('迭代次数') ylabel('准确率') xlim([1, length(info.TrainingAccuracy)]) grid %% 绘制训练loss损失图 figure plot(1: length(info.TrainingLoss), info.TrainingLoss, 'r-', 'LineWidth', 1) legend('TCN训练loss') xlabel('迭代次数') ylabel('loss') xlim([1, length(info.TrainingLoss)]) grid %% 仿真测试 t_sim1 = predict(net, p_train); t_sim2 = predict(net, p_test); %% 数据反归一化 T_sim1 = vec2ind(t_sim1); T_sim2 = vec2ind(t_sim2); %% 准确率误差 acc1 = sum((T_sim1 == T_train)) / M * 100 ; acc2 = sum((T_sim2 == T_test )) / N * 100 ; disp(['训练集准确率为： ', num2str(acc1)]) disp(['测试集准确率为： ', num2str(acc2)]) %% 数据排序 [T_train, index_1] = sort(T_train); [T_test , index_2] = sort(T_test ); T_sim1 = T_sim1(index_1); T_sim2 = T_sim2(index_2); %% 绘图 figure plot(1: M, T_train, 'r-*', 1: M, T_sim1, 'b-o', 'LineWidth', 1) legend('真实值', 'TCN预测值') xlabel('预测样本') ylabel('预测结果') string = {'训练集预测结果对比'; ['准确率=' num2str(acc1)]}; title(string) xlim([1, M]) grid figure plot(1: N, T_test, 'r-*', 1: N, T_sim2, 'b-o', 'LineWidth', 1) legend('真实值', 'TCN预测值') xlabel('预测样本') ylabel('预测结果') string = {'测试集预测结果对比'; ['准确率=' num