<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">On 11/4/2012 12:52 PM, Md. Akmal Haidar
wrote:<br>
</div>
<blockquote
cite="mid:1352062325.43159.YahooMailNeo@web161001.mail.bf1.yahoo.com"
type="cite">
<div style="color:#000; background-color:#fff; font-family:times
new roman, new york, times, serif;font-size:12pt">
<div><span>Thanks Anand,</span></div>
<div style="color: rgb(0, 0, 0); font-size: 16px; font-family:
times new roman,new york,times,serif; background-color:
transparent; font-style: normal;">It works now.<br>
<span></span></div>
<div style="color: rgb(0, 0, 0); font-size: 16px; font-family:
times new roman,new york,times,serif; background-color:
transparent; font-style: normal;"><span>But I found the same
WER for the original n-best list and the rescored nbest
list.</span></div>
<div style="color: rgb(0, 0, 0); font-size: 16px; font-family:
times new roman,new york,times,serif; background-color:
transparent; font-style: normal;"><span><br>
</span></div>
<div style="color: rgb(0, 0, 0); font-size: 16px; font-family:
times new roman,new york,times,serif; background-color:
transparent; font-style: normal;"><span>For rescoring, I use
the following command:</span></div>
<div style="color: rgb(0, 0, 0); font-size: 16px; font-family:
times new roman,new york,times,serif; background-color:
transparent; font-style: normal;"><span>rescore-decipher
nbestfilelist new_rescored_nbestlist_dir -lm updated_lm</span></div>
</div>
</blockquote>
<br>
nbest-error computes the "N-best error rate", meaning the best
possible error rate that can be achieved by picking a hypothesis
from anywhere among the N best. This is sometimes called the
"oracle" error. It doesn't changed as a result of rescoring.<br>
<br>
What you probably want is the error rate of the highest-scoring
hypothesis. For this, you first extract the highest-scoring hyps
using the "rescore-reweight" command (see nbest-scripts(1) man
page), then you using your favorite WER scoring program. If you
have NIST sclite installed, you could use the compute-sclite
wrapper, which takes care of format differences.<br>
<br>
Andreas<br>
<br>
</body>
</html>