Model Validation Using Mutated Training Labels: An Exploratory Study
This addresses the problem of reliable model validation for supervised learning when validation sets are unavailable, offering a novel approach that is incremental in its application to existing methods.
The paper tackles model validation without a separate validation set by introducing Mutation Validation (MV), which mutates training labels and uses performance changes to assess model fit; results show MV achieves a 92% hit rate in model selection, outperforming out-of-sample-validation at less than 60%.
We introduce an exploratory study on Mutation Validation (MV), a model validation method using mutated training labels for supervised learning. MV mutates training data labels, retrains the model against the mutated data, then uses the metamorphic relation that captures the consequent training performance changes to assess model fit. It does not use a validation set or test set. The intuition underpinning MV is that overfitting models tend to fit noise in the training data. We explore 8 different learning algorithms, 18 datasets, and 5 types of hyperparameter tuning tasks. Our results demonstrate that MV is accurate in model selection: the model recommendation hit rate is 92\% for MV and less than 60\% for out-of-sample-validation. MV also provides more stable hyperparameter tuning results than out-of-sample-validation across different runs.