Kernel Ridge Regression Model

This example shows how to train a mixed model with KRR and tree components. We reproduce an experiment from [1] using the wine quality dataset. On this dataset we see that alternating kernel ridge regression and tree models achieves lower test error than either of these models alone.

[1] Sigrist, Fabio. “KTBoost: Combined kernel and tree boosting.” Neural Processing Letters 53.2 (2021): 1147-1160.