Validation Set

The validation set is a portion of data set aside during the model development process to tune hyperparameters and provide an unbiased evaluation of the model before final testing. Unlike the training set, which the model uses to learn, the validation set is used to guide the development and prevent overfitting.

In the context of derivatives, this might involve adjusting the sensitivity of a delta-hedging algorithm to see how it performs on a held-out sample of volatility data. By iteratively checking performance against the validation set, researchers can refine the model to ensure it is not just fitting the training data.

This is a standard procedure in machine learning that is essential for ensuring that the resulting model is both effective and reliable.