eroor in lstm training network

Question

bhanu kiran vandrangi 2022-3-25

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1680414-eroor-in-lstm-training-network

评论： KSSV 2022-3-28

EUR_USD Historical Data.csv

'The training sequences are of feature dimension 1 but the input layer expects sequences of feature dimension 2.' this is the error i am getting whwn i run the code

clear all; clc;
Data=xlsread('EUR_USD Historical Data.csv')
Stock=Data(:,2)
percentage=Data(:,5)*100
[row col]=size(Data)
%create database for lstm
SequenceLength=30
Database={}
resultDatabase=[]
count=0
for i=row:-1:SequenceLength
    count=count+1
    a=Stock(i-SequenceLength+1:i)'
    a=flip(a)
    b=percentage(i-SequenceLength+1:i)'
    b=flip(b)
    Database{count,1}=[a,b]
end
[rt ct]=size(Database)
count=0
for i=1:rt-1
    count=count+1
    a=Database{count+1,1}
    resultDatabase(count,1)=a(1,end)
end 
EURUSDDatabase=Database
save EURUSDDatabase
%lstm creation
for i=1:rt-20
    stockTrain{i,1}=Database{i,1}
end
restTrain=resultDatabase(1:rt-20)
numFeatures=2
numHiddenLayers=125
numResponses=1
Layers=[ sequenceInputLayer(numFeatures),...
    lstmLayer(numHiddenLayers,'OutputMode','Last'),...
    fullyConnectedLayer(numResponses),...
    regressionLayer
    ]
miniBatchSize=3;
Epoch=20;
options=trainingOptions('adam',...
    'ExecutionEnvironment','auto',...
    'MaxEpochs',Epoch,...
    'MiniBatchSize',miniBatchSize,...
    'plots','training-progress')
TradernetEURUSD=trainNetwork(stockTrain,restTrain,Layers,options)
save TradernetEURUSD