Related Experiment Videos
Explained variation and predictive accuracy in general parametric statistical models: the role of model
Susanne Rosthøj1, Niels Keiding
1Department of Biostatistics, University of Copenhagen, Blegdamsvej 3, DK-2200 Copenhagen N, Denmark.
Lifetime Data Analysis
|February 5, 2005
Summary
This study introduces explained variation and predictive accuracy measures for regression models. It details estimation procedures and explores misspecification effects, extending concepts to survival analysis.
Area of Science:
- Statistics
- Biostatistics
- Epidemiology
Background:
- Regression models are crucial for understanding relationships between covariates and outcomes.
- Assessing model performance relies on measures of explained variation and predictive accuracy.
- Existing methods may not fully address model misspecification or survival data.
Purpose of the Study:
- To provide a comprehensive introduction to explained variation and predictive accuracy measures in regression.
- To present a general framework for estimating these measures and studying misspecification.
- To extend these concepts to survival analysis.
Main Methods:
- Detailed theoretical exposition of explained variation and predictive accuracy.
- Development of estimation procedures for these measures.
- Framework for analyzing the impact of model misspecification.
- Generalization of methods for survival data.
Main Results:
- A unified framework for understanding explained variation and predictive accuracy.
- Insights into how model misspecification affects estimated quantities.
- Demonstration of applicability to survival analysis.
Conclusions:
- The presented measures and framework enhance the evaluation of regression models.
- Understanding misspecification is critical for reliable covariate-outcome assessments.
- The generalization to survival analysis broadens the utility of these statistical tools.