Text normalization in NLP vs in ASR.

on

 It is traditional in NLP to separate the 's

In ASR we normally leave it part of the word.

corpus.txt

Musharraf's Last Act?
Normalized text in NLP

Musharraf 's Last Act ?

One way of doing it in ASR. 

musharraf's last act
Can use capital letters if needed. 

MUSHARRAF'S LAST ACT
The main idea is 's is not separated from the word.

Comments

Post a Comment