Calibration and validation in ANN model

I have 3348 data, of which I want I want to use 2232 as calibration data to validate in 1116 data in ANN model. I earlier used 70% training ratio, 15% as validation ratio and 15% as test ratio.

% net.divideParam.trainRatio = 60; % net.divideParam.valRatio = 20; % net.divideParam.testRatio = 20; [net,tr] = train(net,Inputs,targets); nntraintool

Is this is the right way or is there any way? Really appreciated your time and help.

[Nakul]

Did you get a solution?

Answer 1

[Tejas]

Hello Sonisa,

The ‘divideParam’ function expects the ‘trainRatio’, ‘valRatio’, and ‘testRatio’ to be decimal values. Here is the correct code snippet:

net.divideParam.trainRatio = 60/100;
net.divideParam.valRatio = 20/100;
net.divideParam.testRatio = 20/100;

Refer to this documentation for more information:https://www.mathworks.com/help/deeplearning/gs/fit-data-with-a-neural-network.html#:~:text=Set%20up%20the%20division%20of%20data