Theory of Generalization-6 Machine Learning Foundations
Review
break point: minimum data could be shattered
shattered: could represent all dichotomies
dichotomies: number of set could bi-separate the data
m(N): growth function, represent maximum of dichotomies given N
Term
Bounding Function B(N, k): N個點中,任k個不能shattered combinatorial quantity: Maximum number of length N vectors with (o,x), while no shatter in any length ksubvectors
No comments:
Post a Comment