Skip to content

Latest commit

 

History

History
21 lines (17 loc) · 708 Bytes

Section 5-1 to 5-4.md

File metadata and controls

21 lines (17 loc) · 708 Bytes

5.1 Application for spelling correction

用到:最小编辑距离, 候选集

(pronunciation candidate set, spelling candidate set)

5.2 Noisy Channel Intuition

噪声信道的可信度:贝叶斯规则的最大概率估计
几种纠正方法:删除,替换,插入,交换
mispelling tests set

5.3 Real-Word Spelling errors

candidate set and its sequence to maximize the P(W)

5.4 State of the Art Systems

commensurate
1.改进的概率假设:词语概率非条件独立
2.factors that could influence p(mispelling | word)
source word, target word
3.
除了channel model和language model,还有其它的model
概率分布计算的模型