How to estimate by already trained network

I utilized ANN-GA to train a network in which X is input and Y is output. Now I want to use the trained network to estimated new Y data (1column in excel) by new X data ( 6 columns in excel). I don't now the code. Please help me to write the code. I also attached the code.

%% Data Loading

Data = xlsread('pb_2.xlsx');

X = Data(:,1:end-1);

Y = Data(:,end);

DataNum = size(X,1);

InputNum = size(X,2);

OutputNum = size(Y,2);

%% Normalization

MinX = min(X);

MaxX = max(X);

MinY = min(Y);

MaxY = max(Y);

XN = X;

YN = Y;

for ii = 1:InputNum

XN(:,ii) = Normalize_Fcn(X(:,ii),MinX(ii),MaxX(ii));

end

for ii = 1:OutputNum

YN(:,ii) = Normalize_Fcn(Y(:,ii),MinY(ii),MaxY(ii));

end

%% Test and Train Data

TrPercent = 80;

TrNum = round(DataNum * TrPercent / 100);

TsNum = DataNum - TrNum;

R = randperm(DataNum);

trIndex = R(1 : TrNum);

tsIndex = R(1+TrNum : end);

Xtr = XN(trIndex,:);

Ytr = YN(trIndex,:);

Xts = XN(tsIndex,:);

Yts = YN(tsIndex,:);

%% Network Structure

pr = [-1 1];

PR = repmat(pr,InputNum,1);

Network = newff(PR,[5 OutputNum],{'tansig' 'tansig'});

%% Training

Network = TrainUsing_GA_Fcn(Network,Xtr,Ytr);

%% Assesment

YtrNet = sim(Network,Xtr')';

YtsNet = sim(Network,Xts')';

MSEtr = mse(YtrNet - Ytr)

MSEts = mse(YtsNet - Yts)

RMSEtr= sqrt(MSEtr)

RMSEts= sqrt(MSEts)

Bias_tr= (sum(Ytr- YtrNet))/ length(Ytr)

Bias_ts= (sum(Yts- YtsNet))/ length(Yts)

SSEtr= sum((Ytr - YtrNet).^2);

SSEts= sum((Yts - YtsNet).^2);

SSTtr= sum((Ytr - mean(Ytr)).^2);

SSTts= sum((Yts - mean(Yts)).^2);

Rsquare_tr= (sum(Ytr .* YtrNet)) / (sqrt(sum(Ytr .^ 2)) * (sum(YtrNet .^ 2)))

Rsquare_ts= (sum(Yts .* YtsNet)) / (sqrt(sum(Yts .^ 2)) * (sum(YtsNet .^ 2)))

Answer 1

[Purvaja]

Hi @Arman Naderi,

I understand that you want to estimate new Y values. As seen from your given code, “YtsNet” variable is storing the estimated results after the model is trained on “Xts” data from “pb_2.xls” sheet, that you loaded initially in the code.

1. If you wish to store the new estimated values in excel sheet alongside X values, you may try following steps:

%% Estimated data
YtsNet = sim(Network, Xts')'; 

% Create a column vector of NaNs for all rows  
EstimatedY = nan(DataNum, 1); 

% Insert predictions for the test set rows  
EstimatedY(tsIndex) = YtsNet; 
DataWithEstimatedY = [Data EstimatedY]; 
xlswrite('pb_2_with_estimated.xlsx', DataWithEstimatedY); 

2. If you wish to store the trained model and then load to model to use it on different dataset, you may try following steps:

%% Training the network  
Network = TrainUsing_GA_Fcn(Network, Xtr, Ytr);  

%% Save the trained network and normalization parameters  
save('trainedModel.mat', 'Network'); 

In another file you can:

%% Load the trained model and normalization parameters 
load('trainedModel.mat');  

%% Load new input data (assume it has the same number of columns as X) 
NewData = xlsread('new_data.xlsx'); 

%% Predict using the loaded model  
NewY = sim(Network, NewXN')'; 

You can also go through following documentation link to know more about “load” function:

https://www.mathworks.com/help/matlab/ref/load.html

Hope this helps you!