Error for input into Neural Network

3 次查看（过去 30 天）

显示更早的评论

Killian Flynn 2022-12-1

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1868623-error-for-input-into-neural-network

编辑： Killian Flynn 2022-12-7

EURUSD=X.csv

在 MATLAB Online 中打开

Hello. I am trying to train a neural network but I am getting an error. My code is:

data =readtable('EURUSD=X.csv')
[m,n] = size(data);
k = floor(m/20)*20; % gives me exact number of rows of data for 20 days period
data = data(1:k,:); % removes extra unwanted trailing data
Open = data(:,2);
High= data(:,3);
Low = data(:,4);
Close = data(:,5);
%Store the data Open,High, Low and Close for every 20 days in the variable
%DataE20
G = floor((0:height(data)-1)/20).' + 1;
DataE20 = splitapply(@(Date,Open,High,Low,Close,AdjClose,Volume) {[Open,High,Low,Close].'}, data, G)
celldisp(DataE20(13))
Start = cellfun(@(x)x(1),DataE20)
Finish = cellfun(@(x)x(end),DataE20)
x = length(DataE20)
for y = 1:x
   if Finish(y)>Start(y)
       Labels(y) = 1
    else
        Labels(y) = 0  
    end
end
  
for layer1 = 1:15
    for layer2 = 1:15
        network = patternnet([layer1 layer2]);
        network.trainParam.showWindow = false;
        trained_net = train(network,DataE20,Labels); %network,input,target
        op = trained_net(DataE20);
        conf = confusion(Labels,op);
        if best_conf>conf
            best_conf = conf;
            best_net = trained_net;
        end
    end
end
best_conf

I get the error:

Error using network/train (line 340)

Number of inputs does not match net.numInputs.

Error in Project1 (line 56)

trained_net = train(network,DataE20,Labels); %network,input,target

How can I fix this? Any help would be greatly appreciated.

2 个评论
显示无隐藏无

Sudarshan 2022-12-6

Could you let me know why you are passing DataE20 as a cell array of 4x20 doubles into the input?

Killian Flynn 2022-12-7

编辑：Killian Flynn 2022-12-7

Hello. Originally I thought I could give each 4x20 doubles a label but it seems as though it needs to be 1x80 instead. The idea is to train the AI so that it take the data from every 20 days and uses it to predict whether the next 20 days will be up or down where up is a 1 and down is a 0. That's the general idea. The data is made up of the Open, close low and high which I thought would make the rows and the data for every 20 days to make the colmns. I hope that helps.

请先登录，再进行评论。

请先登录，再回答此问题。