Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
August 10, 2026Econometric TheoryOpen Access

On Asymptotic Optimality of Least Squares Model Averaging When True Model Is included

View Full Paper
Ask AI
Bookmark
Share

Authors

WXWenchao XuXZXinyu Zhang

Discussion

Loading...

Member takes

Overview

Randomized trial shows asymptotic properties of model averaging methods, indicating different outcomes with varying model dimensions.

Key Points

  • This study aims to address technical challenges in model averaging by analyzing least squares methods with nested candidate models.
  • Investigated least squares model averaging methods including Mallows model averaging and parsimonious model averaging.
  • Compared asymptotic properties under fixed vs. diverging true model dimensions.
  • Conducted simulation studies for theoretical validation.
  • Mallows model averaging with fixed dimensions is neither loss optimal nor risk optimal.
  • When true model dimension diverges, Mallows model averaging is both loss and risk optimal.
  • Parsimonious model averaging is risk optimal but not loss optimal when the true model dimension is fixed.

Cite This Study

Xu et al. (2026) studied this question.

synapsesocial.com/papers/6a797ce09c20a9bbd3183e86https://doi.org/10.1017/s0266466626100553
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Optimal model averaging for partially linear models with missing response variables and error‐prone covariates2024 · 1 citations
  2. 2Model Average Estimation of Parameters in Linear Model With Multiple Change Points2025
  3. 3Model Averaging for Accelerated Failure Time Models with Missing Censoring Indicators2024
  4. 4Geometric Model Averaging2026
  5. 5Model Averaging of Partially Linear Multinomial Logit Model for Multi‐Categorical Data2026