【LSTM时序预测】基于卷积神经网络结合BiLSTM实现时序数据预测附matlab代码

⛄ 内容介绍

潮汐电站的优化运行需要进行潮位预测.针对传统调和分析方法不能有效处理潮位非线性和非平稳的特性的问题,提出一种CNN-BiLSTM的预测模型,以滑动数据窗口构造潮位数据的特征图作为输入,利用1D CNN提取潮位数据深层特征,然后采用BiLSTM网络生成特征描述,最后输出预测结果.采用芝加哥港口的潮汐数据进行了实验,所提预测模型与调和分析及LSTM模型相比,均方根误差分别降低了66.26%和30.11%.算例结果表明CNN-BiLSTM模型可以实现高精度的短期潮位预测.

⛄ 部分代码

clc;clear all;close all;

%synthetic reach data from LQG 2 joint controller
load('dat3.mat')
load netIMtoAng.mat
load netAngToPos.mat

YPred = predict(netIMtoAng,dat3);
Out1  = netAngToPos(YPred');

%% visualise  decoded data
for i= 1:160
    % plot path in joint space
    p1 =  Out1(1:51,i);
    p2 =  Out1(52:102,i);
    v1 =  Out1(103:153,i);
    v2 =  Out1(154:204,i);
    subplot(131);plot(p1,p2,'.k'); title('Pos');hold on
    subplot(132);plot(v1,v2,'.k'); title('Vel');
    hold on;
%     pause
    subplot(133); plot(tTra(3,i),tTra(4,i),'o'); hold on;
end
% axis image;
subplot(132);xlabel('shoulder (rad)');
ylabel('elbow (rad)');

%% Test
t = Tiff('M7.tif','r');
S1   = read(t);
s{7} = rgb2gray(S1);
B{7} = imresize(s{7},[28 28]);
TestImage(:,:,1,1) = B{7}(:,:);

YPred = predict(netIMtoAng,TestImage(:,:,1,1));
Out2  = netAngToPos(YPred');

subplot(141); imshow(S1);title('Cartesian Work space')
subplot(142); plot(tTra(3,7),tTra(4,7),'or'); hold on;
plot(pi/2,pi/2,'or'); hold on;
xlim([0.8,2.4]);ylim([0.8,2.4]);axis square;title('Joint space');
xlabel('shoulder (rad)');
subplot(143);
p1 =  Out2(1:51,1);
p2 =  Out2(52:102,1);
v1 =  Out2(103:153,1);
v2 =  Out2(154:204,1);
plot(p1,p2,'.k'); title('Pos');hold on;
xlim([0.8,2.4]);ylim([0.8,2.4]);
xlabel('shoulder (rad)');
ylabel('elbow (rad)');
axis square
subplot(144);plot(v1,v2,'.k'); title('Vel');
xlabel('shoulder (rad/s)');
ylabel('elbow (rad/s)');
xlim([-2.5,2.5]);ylim([-2.5,2.5]);
axis square

%% Lstm
data  = Out1(3:51,1)';
% dataY =  Out1(54:102,1)';
% velX  =  Out1(105:153,1)';
% velY  =  Out1(156:204,1)';
% DD    = [dataX; dataY; velX; velY];

figure
plot(data)

%% prepare input -output for lstm
numTimeStepsTrain = floor(0.9*numel(data));

dataTrain = data(1:numTimeStepsTrain+1);
dataTest  = data(numTimeStepsTrain+1:end);
mu        = mean(dataTrain);
sig       = std(dataTrain);
dataTrainStandardized = (dataTrain - mu) / sig;
XTra = dataTrainStandardized(1:end-1);
YTra = dataTrainStandardized(2:end);

%% LSTM training
numFeatures = 1;
numResponses = 1;
numHiddenUnits = 200;

layers = [ ...
    sequenceInputLayer(numFeatures)
    lstmLayer(numHiddenUnits)
    fullyConnectedLayer(numResponses)
    regressionLayer];

options = trainingOptions('adam', ...
    'MaxEpochs',250, ...
    'GradientThreshold',1, ...
    'InitialLearnRate',0.005, ...
    'LearnRateSchedule','piecewise', ...
    'LearnRateDropPeriod',125, ...
    'LearnRateDropFactor',0.2, ...
    'Verbose',0, ...
    'Plots','training-progress');

net = trainNetwork(XTra,YTra,layers,options);
analyzeNetwork(net)
%% forcast- predicted trajectory
dataTestStandardized = (dataTest - mu) / sig;
XTest = dataTestStandardized(1:end-1);
net = predictAndUpdateState(net,XTra);
[net,YPred] = predictAndUpdateState(net,YTra(end));

numTimeStepsTest = numel(XTest);
for i = 2:numTimeStepsTest
    [net,YPred(:,i)] = predictAndUpdateState(net,YPred(:,i-1),'ExecutionEnvironment','cpu');
end
YPred = sig*YPred + mu;
YTest = dataTest(2:end);
rmse = sqrt(mean((YPred-YTest).^2))
figure
plot(dataTrain(1:end-1),'Linewidth',6)
hold on
idx = numTimeStepsTrain:(numTimeStepsTrain+numTimeStepsTest);
plot(idx,[data(numTimeStepsTrain) YPred],'.-','Markersize',32)
hold off
legend(['Observed', 'Forecast']);
set(gca,'Fontsize',20);
xlabel('Time steps');
ylabel('Shoulder angle position(rad)')

⛄ 运行结果

⛄ 参考文献

[1]黄冬梅王唱胡安铎孙锦中孙园李俊峰. 基于CNN-BiLSTM的潮汐电站潮位预测[J]. 水力发电, 2021, 047(010):80-84.

[2]金啸宇. 基于LSTM神经网络的多元时间序列预测[D]. 东北大学.​

⛄ Matlab代码关注

❤️部分理论引用网络文献，若有侵权联系博主删除

❤️ 关注我领取海量matlab电子书和数学建模资料