Bounding the Vapnik-Chervonenkis Dimension of Concept Classes Parameterized by Real Numbers

Machine Learning

Volumn 18, Issue 2, 1995, Pages 131-148

  Goldberg, Paul W ^a   Jerrum, Mark R ^a  

^a NONE

Author keywords

Concept learning; information theory; Milnor's theorem; Vapnik Chervonenkis dimension

Indexed keywords

COMBINATORIAL MATHEMATICS; COMPUTATIONAL COMPLEXITY; DIGITAL ARITHMETIC; GEOMETRY; INFORMATION THEORY; NEURAL NETWORKS; POLYNOMIALS;

INDUCTIVE CONCEPT LEARNING; MILNOR'S THEOREM; OCCAM SENSE; POLYNOMIAL LEARNABILITY; PROBABLY APPROXIMATELY CORRECT FRAMEWORK; REAL NUMBERS; SYNTACTIC COMPLEXITY; VAPNIK-CHERVONENKIS DIMENSION;

LEARNING SYSTEMS;

EID: 0029256399     PISSN: 08856125     EISSN: 15730565     Source Type: Journal    
DOI: 10.1023/A:1022847329178     Document Type: Article

References (29)

1. A Linear Time Algorithm for the Hausdorff-distance between Convex Polygons
2. What Size Net Gives Valid Generalization?
3. Occam's Razor
4. Learnability and the Vapnik-Chervonenkis Dimension
5. Central Limit Theorems for Empirical Measures
6. A General Lower Bound on the Number of Examples Needed for Learning
7. Vapnik-Chervonenkis Classes of Definable Sets
8. Results on Learnability and the Vapnik-Chervonenkis Dimension
9. On the Betti Numbers of Real Varieties
10. On the Computational Complexity and Geometry of the First-Order Theory of the Reals
11. Feedforward Nets for Interpolation and Classification
12. Lower Bounds for Algebraic Decision Trees
13. A Theory of the Learnable
14. A View of Computational Learning Theory
15. On the uniform convergence of relative frequencies of events to their probabilities
16. Lower Bounds for Approximation by Non-linear Manifolds
17. Some special Vapnik-Chervonenkis classes