Optimal design of regularization term and regularization parameter by subspace information criterion

Neural Networks

Volumn 15, Issue 3, 2002, Pages 349-361

  Sugiyama, Masashi ^a   Ogawa, Hidemitsu ^a  

^a TOKYO INSTITUTE OF TECHNOLOGY   (Japan)

Author keywords

Generalization error; Linear regression; Model selection; Regularization learning; Regularization parameter; Ridge regression; Subspace information criterion; Supervised learning

Indexed keywords

APPROXIMATION THEORY; COMPUTER SIMULATION; DATA PROCESSING; ERRORS; OPTIMIZATION;

REGULARIZATION TERMS;

NEURAL NETWORKS;

ARTICLE; COMPUTER SIMULATION; INFORMATION; LEARNING; LINEAR REGRESSION ANALYSIS; PAIRED ASSOCIATE LEARNING; PRIORITY JOURNAL; STATISTICAL MODEL; THEORETICAL MODEL;

LINEAR MODELS; MODELS, THEORETICAL;

EID: 0035989166     PISSN: 08936080     EISSN: None     Source Type: Journal    
DOI: 10.1016/S0893-6080(02)00022-9     Document Type: Article

References (24)

1. A new look at the statistical model identification
2. Likelihood and the Bayes procedure
3. Model complexity control for regression using VC generalization bounds
4. Smoothing noisy data with spline functions: Estimating the correct degree of smoothing by the method of generalized cross-validation
5. Generalized information criterion in model selection
6. Iterative choices of regularization parameters in linear inverse problem
7. Bayesian interpolation
8. Choosing variables in a linear regression: A graphical aid
9. P
10. Bias of estimators and regularization terms
11. Network information criterion - Determining the number of hidden units for an artificial neural network model
12. Statistical aspects of model selection
13. Subspace information criterion for model selection
14. Theoretical and experimental evaluation of subspace information criterion. Special Issue on New Methods for Model Selection and Model Combination
15. Algebraic analysis for non-identifiable learning machines