Cross-validation Approaches for Multi-study Predictions

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ren, Boyu, Patil, Prasad, Dominici, Francesca, Parmigiani, Giovanni, Trippa, Lorenzo
Format: Preprint
Published: 2020
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916330529095680
author Ren, Boyu
Patil, Prasad
Dominici, Francesca
Parmigiani, Giovanni
Trippa, Lorenzo
author_facet Ren, Boyu
Patil, Prasad
Dominici, Francesca
Parmigiani, Giovanni
Trippa, Lorenzo
contents We consider prediction in multiple studies with potential differences in the relationships between predictors and outcomes. Our objective is to integrate data from multiple studies to develop prediction models for unseen studies. We propose and investigate two cross-validation approaches applicable to multi-study stacking, an ensemble method that linearly combines study-specific ensemble members to produce generalizable predictions. Among our cross-validation approaches are some that avoid reuse of the same data in both the training and stacking steps, as done in earlier multi-study stacking. We prove that under mild regularity conditions the proposed cross-validation approaches produce stacked prediction functions with oracle properties. We also identify analytically in which scenarios the proposed cross-validation approaches increase prediction accuracy compared to stacking with data reuse. We perform a simulation study to illustrate these results. Finally, we apply our method to predicting mortality from long-term exposure to air pollutants, using collections of datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2007_12807
institution arXiv
publishDate 2020
record_format arxiv
spellingShingle Cross-validation Approaches for Multi-study Predictions
Ren, Boyu
Patil, Prasad
Dominici, Francesca
Parmigiani, Giovanni
Trippa, Lorenzo
Methodology
Statistics Theory
We consider prediction in multiple studies with potential differences in the relationships between predictors and outcomes. Our objective is to integrate data from multiple studies to develop prediction models for unseen studies. We propose and investigate two cross-validation approaches applicable to multi-study stacking, an ensemble method that linearly combines study-specific ensemble members to produce generalizable predictions. Among our cross-validation approaches are some that avoid reuse of the same data in both the training and stacking steps, as done in earlier multi-study stacking. We prove that under mild regularity conditions the proposed cross-validation approaches produce stacked prediction functions with oracle properties. We also identify analytically in which scenarios the proposed cross-validation approaches increase prediction accuracy compared to stacking with data reuse. We perform a simulation study to illustrate these results. Finally, we apply our method to predicting mortality from long-term exposure to air pollutants, using collections of datasets.
title Cross-validation Approaches for Multi-study Predictions
topic Methodology
Statistics Theory
url https://arxiv.org/abs/2007.12807