Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Chen, Ke, Jiang, Dandan, Zhang, Xinyu
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2605.13203
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866916070316572672
author Chen, Ke
Jiang, Dandan
Zhang, Xinyu
author_facet Chen, Ke
Jiang, Dandan
Zhang, Xinyu
contents This paper investigates the predictive performance of model averaging in high-dimensional linear regression where the number of regressors is comparable to the sample size. Leveraging tools from random matrix theory, we derive the exact limiting out-of-sample risk under a nested model setting and comprehensively characterize the risk landscape. This limiting risk helps to reveal two phenomena: simple weighting inherits the double descent trajectory and its associated variance explosion near the interpolation boundary; strategic weighting triggers an ensemble emergence that suppresses the localized risk surge and yields a globally flat risk surface. Building on this limiting risk, we also propose the Large Model Averaging (LaMA) method, in which we consider the discrepancy between in-sample and out-of-sample risks in the high-dimensional regime. Numerical studies and real data applications confirm that LaMA achieves superior predictive accuracy in high-dimensional environments.
format Preprint
id arxiv_https___arxiv_org_abs_2605_13203
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Double Descent and Ensemble Emergence in Model Averaging Prediction
Chen, Ke
Jiang, Dandan
Zhang, Xinyu
Methodology
This paper investigates the predictive performance of model averaging in high-dimensional linear regression where the number of regressors is comparable to the sample size. Leveraging tools from random matrix theory, we derive the exact limiting out-of-sample risk under a nested model setting and comprehensively characterize the risk landscape. This limiting risk helps to reveal two phenomena: simple weighting inherits the double descent trajectory and its associated variance explosion near the interpolation boundary; strategic weighting triggers an ensemble emergence that suppresses the localized risk surge and yields a globally flat risk surface. Building on this limiting risk, we also propose the Large Model Averaging (LaMA) method, in which we consider the discrepancy between in-sample and out-of-sample risks in the high-dimensional regime. Numerical studies and real data applications confirm that LaMA achieves superior predictive accuracy in high-dimensional environments.
title Double Descent and Ensemble Emergence in Model Averaging Prediction
topic Methodology
url https://arxiv.org/abs/2605.13203