N-grams model

针对OOV两种方式
cs224n Lecture6 Language Models and RNNs
稀疏问题:
1.分子为0
使用Smoothing(discounting)

  • Laplace smoothing(add-1 smoothing):
    discount dc
    cs224n Lecture6 Language Models and RNNs
    cs224n Lecture6 Language Models and RNNs

cs224n Lecture6 Language Models and RNNs

  • Add-k smoothing:
    cs224n Lecture6 Language Models and RNNs
    2.分母为0

  • Backoff and Interpolation:
    we only “back off” to a lower-order n-gram if we have zero evidence for a higher-order n-gram
    held-out 、discount
    cs224n Lecture6 Language Models and RNNs

  • Katz backoff
    cs224n Lecture6 Language Models and RNNs

  • Kneser-Ney Smoothing

PERPLEXITY’S RELATION TO ENTROPY

  • Entropy
    cs224n Lecture6 Language Models and RNNs
  • Entory rate

cs224n Lecture6 Language Models and RNNs

相关文章: