EM-ELM Cross validation

Question

I know that cross validation is used to find the best hyperparameters that minimize the average error. For example, the number of neurons that minimize the average error of cross-validation is estimated in ELM.

But I would like to know how I can apply cross-validation (eg K-Fold) in EM-ELM networks:

Feng, G., Huang, G. B., Lin, Q., & Gay, R. (2009). Error minimized extreme learning machine with growth of hidden nodes and incremental learning. IEEE Transactions on Neural Networks, 20 (8), 1352-1357.

Where the architecture is estimated automatically using a training set. What is the best and most standard way to apply cross-validation in EM-ELM? and is it possible to incorporate the cross-validation process into the growth of the network?

Answer 1

Cross-validation can be used to tune any hyperparameter.

One way tune Extreme learning machines (ELM) with cross-validation is to:

1. Fit a model to a subset of the training dataset. Measure performance on a validation dataset with a evaluation metrics.
2. Change the number of neurons hyperparameter.
3. Fit the updated model to a subset of the training dataset. Measure performance on a validation dataset with a evaluation metrics.
4. Compare the two models and see which model is better on a evaluation metric on their respective validation dataset.
5. Repeat steps 3-4 until satisfied.