could anyone help me how to validate the neural network

I am developing neural network model to train, validate and test the network.

For training I am using net = trainNetwork(XTrain,YTrain,layers)

For testing I am using YPred = predict(net,XTest).

Before testing I need to do validation with the trained network using validation XVal.

Could anyone please help me how to do validation for the trained network before testing.

[Image Analyst]

Normally you have 3 sets of data:

1. A training set with known ground truth values. This is used in the training process.
2. A validation set with known ground truth values. This is used in the training process to decide when to quit training.
3. A test set with unknown values.

It's up to you to collect your 3 sets of data. What happened to your validation set? Did you not use one?

[jaah navi]

I have collected training dataset, validation dataset and testing dataset.

Also,I am not using tool box, I am writing the code my myself.

In my code, for the training i am using the command line net = trainNetwork(XTrain,YTrain,layers) to train the model.

Once training is done i need to validate the model using validation dataset(XVal,YVal)

Could you please help me how I need to write the command to do validation using (XVal,YVal) with the already trained model.

[Image Analyst]

You wrote your own trainNetwork() function? Well, you should know how to pass data to it then. Sorry I can't help you anymore with your custom function(s) - I prefer to use Toolbox functions rather than write my own.

Answer 1

[KSSV]

Read the documentation of TrainNetwork, you can pass the validation data to the function.