An Experimental and Theoretical Comparison of Model Selection Methods

• 1 January 1997
• journal article
• Published by Springer Nature in Machine Learning
• Vol. 27  (1) , 7-50
• https://doi.org/10.1023/a:1007344726582

Abstract

Keywords

• SIMPLE MODEL SELECTION PROBLEM
• WELL-KNOWN MODEL SELECTIONMETHODS
• MODEL SELECTIONMETHODS
• SMALL GENERALIZATION ERROR ONBOTH
• THEORETICAL COMPARISON
• GENERALIZATION ERROR
• PENALTY-BASED METHOD
• OBSERVED ERROR ONA
• ADDITIONAL ERROR OFCROSS
• MODEL SELECTION PROBLEM
• COMPLEXITY OFTHE
• BOOLEAN FUNCTION
• MODEL SELECTION
• SUPERVISED LEARNING
• CROSS VALIDATION
• STRUCTURAL RISK MINIMIZATION
• VC DIMENSION
• COMPARATIVE METHOD