Coverage-based performance of confidence intervals for linear regression coefficients under multicollinearity: Simulation and application

View Full Paper

Ask AI

Bookmark

Authors

Discussion

Simulation analysis examines confidence interval coverage in ridge regression, highlighting implications for multicollinearity.

Most ridge regression estimators provide narrower confidence intervals than ordinary least squares, yet some do not maintain 95% coverage probability.
As multicollinearity increases, the width of confidence intervals also widens, complicating inference from linear regression models.
Eighteen ridge regression regularization parameters were compared using a simulation approach, focusing on coverage probability and width.
The study highlights the limitation of mean square error as a measure for inferential performance in ridge regression models.

Chowdhury et al. (2025) studied this question.

View Full Paper

Ask AI

Bookmark