An information approach to regularization parameter selection under model misspecification

Keyphrases

• Regularization Parameter Selection 100%
• Model Misspecification 100%
• Information Approach 100%
• Selection Strategy 40%
• Gaussian Noise 40%
• Information Complexity 40%
• Expected Log-likelihood 40%
• Engineering Application 20%
• Parameter Values 20%
• Linear Relationship 20%
• Model Distribution 20%
• Ridge Regression 20%
• Approximate Model 20%
• Predictive Error 20%
• Regularization Parameter 20%
• Information Criteria 20%
• Discrete Ill-posed Problem 20%
• Maximum Penalized Likelihood 20%
• Distributional Misspecification 20%
• Takeuchi Information Criterion 20%
• Kullback-Leibler Information 20%
• Penalized Likelihood 20%
• Complexity Regularization 20%
• Regularized Solution 20%

Mathematics

• Regularization 100%
• Model Misspecification 100%
• Log Likelihood 33%
• Random Noise 33%
• Information Criterion 33%
• Posed Problem 16%
• Ridge Regression 16%
• Misspecification 16%
• Linear Relationship 16%