what's the meaning of back-off weights in LM

Speech Recognition Toolkit

Brought to you by: air, arthchan2003, awb, bhiksha, and 5 others

This project can now be found here.

what's the meaning of back-off weights in LM

Forum: Help

Creator: CAO.Z.H

Created: 2007-11-14

Updated: 2012-09-22

CAO.Z.H - 2007-11-14

Sphinx uses ARPA-standard format language model.
In LM, there are mainly two information for every n-gram: probs and back-off weights (bo_wt).
I can understand probs, but i can understand back-off weights.
who can tell me something about how to calculate back-off weights?

Thanks

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- CAO.Z.H - 2007-11-15
  
  Thanks again!
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- hiyassat - 2007-11-14
  
  see this document for Back-off
  http://ssli.ee.washington.edu/people/bilmes/ee516/lecs/lec14_slides.pdf
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Log in to post a comment.