Multi-Study Boosting: Theoretical Considerations for Merging vs. Ensembling
This work addresses cross-study replicability for researchers in fields like bioinformatics, offering incremental theoretical insights for boosting algorithms.
The paper tackles the problem of deciding between merging or ensembling studies when training cross-study replicable prediction models, providing theoretical guidelines based on an analytical transition point and verifying them in simulations and a breast cancer gene expression application.
Cross-study replicability is a powerful model evaluation criterion that emphasizes generalizability of predictions. When training cross-study replicable prediction models, it is critical to decide between merging and treating the studies separately. We study boosting algorithms in the presence of potential heterogeneity in predictor-outcome relationships across studies and compare two multi-study learning strategies: 1) merging all the studies and training a single model, and 2) multi-study ensembling, which involves training a separate model on each study and ensembling the resulting predictions. In the regression setting, we provide theoretical guidelines based on an analytical transition point to determine whether it is more beneficial to merge or to ensemble for boosting with linear learners. In addition, we characterize a bias-variance decomposition of estimation error for boosting with component-wise linear learners. We verify the theoretical transition point result in simulation and illustrate how it can guide the decision on merging vs. ensembling in an application to breast cancer gene expression data.