clc
clear
close all
%读取double格式数据
[num1,ax,ay ]= xlsread('杨15578365004.csv',1);
num1(:,1)=[];
n=size(num1,1);
m =size(num1,2);
for ii = n:-1:1
for jj = 1:m
if isnan(num1(ii,jj))
num1(ii,:) = [];
break
end
end
end
nn = 1500;
mm = randperm(length(num1));
num2 =[];
for ii = 1:50
num2 = [num2;[num1(mm(1:nn),:)' [1:m]']];
end
num=[];
num = [mean(num2(:,1:end-1)')' max(num2(:,1:end-1)')' min(num2(:,1:end-1)')' std(num2(:,1:end-1)')' svd(num2(:,1:end-1)') num2(:,end) ];
m0 = round(0.8*length(num));
n = randperm(length(num));
input_train =num(n(1:m0),1:5);%训练数据输入数据
output_train = num(n(1:m0),6);%训练数据输出数据
input_test = num(n(m0+1:end),1:5);%测试数据输入数据
output_test = num(n(m0+1:end),6);%测试数据输出数据
[inputn,inputps]=mapminmax(input_train',-1,1);%训练数据的输入数据的归一化
[outputn,outputps]=mapminmax(output_train',-1,1);%训练数据的输出数据的归一化de
inputn_test=mapminmax('apply',input_test',inputps);
%% Define Network Architecture
% Define the network architecture.
numFeatures = size(num(:,1:5),2);%输入层维度
numResponses = size(num(:,end),2);%输出维度
% 200 hidden units
numHiddenUnits = 200;%第一层维度
% a fully connected layer of size 50 & a dropout layer with dropout probability 0.5
layers = [ ...
sequenceInputLayer(numFeatures)%输入层
lstmLayer(numHiddenUnits,'OutputMode','sequence')%第一层
fullyConnectedLayer(100)%链接层
dropoutLayer(0.2)%遗忘层
fullyConnectedLayer(numResponses)%链接层
regressionLayer];%回归层
% Specify the training options.
% Train for 60 epochs with mini-batches of size 20 using the solver 'adam'
maxEpochs = 60;%最大迭代次数
miniBatchSize = 100;%最小批量
% the learning rate == 0.01
% set the gradient threshold to 1
% set 'Shuffle' to 'never'
options = trainingOptions('adam', ... %解算器
'MaxEpochs',maxEpochs, ... %最大迭代次数
'MiniBatchSize',miniBatchSize, ... %最小批次
'InitialLearnRate',0.01, ... %初始学习率
'GradientThreshold',inf, ... %梯度阈值
'Shuffle','every-epoch', ... %打乱顺序
'Plots','training-progress',... %画图
'Verbose',0); %不输出训练过程
%% Train the Network
net = trainNetwork(inputn,outputn,layers,options);%开始训练
%% Test the Network
y_pred = predict(net,inputn_test,'MiniBatchSize',100)';%测试仿真输出
y_pred=round(mapminmax('reverse',y_pred,outputps))';
%%
figure%打开一个图像窗口
plot(output_test,'k*')%黑色实线，点的形状为*
hold on%继续画图
plot(y_pred,'bo')%蓝色实线，点的形状^
hold off%停止画图
title('测试图')%标题
ylabel('')%Y轴名称
legend('实际值','测试值')%标签
set(gca,'YTick',1:m)
set(gca,'YTicklabel',{'NH_{3}','NO_{2}'})
ylim([0 m+1])
error1 = y_pred'-output_test;%误差
figure
plot(error1,'k-*')
title('测试误差图')
ylabel('误差')
[MSE,RMSE,MBE,MAE ] =MSE_RMSE_MBE_MAE(output_test',y_pred)
R2 = R_2(output_test',y_pred)
