• Idefisk
  • Tools
  • Tutorials
  • Reviews
  • VoIP Providers
  • Archives
ZOIPER softphone
AsteriskGuru Archives
Mailing List Archives
 
 FAQFAQ   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   RegisterRegister 
 ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in 

[asterisk-speech-rec] Speech Score for Individual Tokens CRM

 
   AsteriskGuru Archives Forum Index -> Asterisk-Speech-Rec
View previous topic :: View next topic  
Author Message
StephenKeller at LumenVox
Guest





PostPosted: Wed Jun 06, 2007 8:25 pm    Post subject: [asterisk-speech-rec] Speech Score for Individual Tokens CRM

John,
 
A couple of things:
 
1. The optional $WordsList at the start of the Tag rule seems unnecessary, and was creating multiple parse trees. I think you can remove it.
 
2. I'm not sure what is going on with $$$.score but it didn't seem to be working correctly. Instead we quickly converted the grammar over to use the latest SISR standard (which we support but which is not very well documented by us yet -- it's high on the to-do list). The grammar below seems to do what you want, you may want to check out the SISR spec to get a better idea of the syntax: http://www.w3.org/TR/semantic-interpretation/
 
Also, as a tip to everybody working with grammars: our Grammar Editor is included as part of the Speech Tuner, which is included with your LumenVox Speech Engine purchase (it's a separate download). It's a very handy tool, as you can parse phrases against a grammar and immediately see exactly how we're parsing it. If anyone has purchased LumenVox and needs the Tuner download link (just note it's a Windows application), feel free to e-mail me.
 
Here's the grammar:
 
#ABNF 1.0;
language en-US;
mode voice;
tag-format <semantics/1.0.2006>;
 
root $FullTagGrammar;
 
$everyone = everyone{out="everyone"};
$having = having{out="having"};
$great = great{out="great"};
$time = time{out="time"};
$holidays = holidays{out="holidays"};
$later = later{out="later"};
$beach = beach{out="beach"};
$house = house{out="house"};
$barbeque = barbeque{out="barbeque"};
$beer = beer|beers{out="beer"};
$see = see{out="see"};
$soon = soon{out="soon"};
 
$WordsList = ($everyone|$having|$great|$time|$holidays|$later|$beach|$house|$barbeque|$beer|$see|$soon){out = rules.latest()};
 
$Tag = $WordsList {out.wordslist = rules.latest(); out.score = meta.latest().score};
 
$FullTagGrammar = {out.tagslist =new Array} ($Tag {out.tagslist.push(rules.latest())} )<1-> [and $Tag{out.tagslist.push(rules.latest())}];
 
 

Stephen Keller
LumenVox Support
P: 877-977-0707, just say "Support"
F: 858-707-7072
Support@LumenVox.com (support@lumenvox.com)
www.LumenVox.com
[img]cid:835062317@06062007-2F12[/img]
Winner “Best Innovation in Speech Recognition”
AVIOS SpeechTEK Award  


Quote:
From: asterisk-speech-rec-bounces@lists.digium.com [mailto:asterisk-speech-rec-bounces@lists.digium.com] On Behalf Of help
Sent: Wednesday, June 06, 2007 7:57 AM
To: asterisk-speech-rec@lists.digium.com
Subject: [asterisk-speech-rec] Speech Score for Individual Tokens



I'm having a problem return the above mentioned. I've mailed someone on this and they mentioned the $$$.score variable which  should return the confidence score of a particular rule in a grammar context. As can be seen below in the sample XML output from my test gramma this seems to be returning undefined. My grammar. I hope someone points out the blindingly obvious and makes me say Doh!! :)


Grammar
=====================================
#ABNF 1.0;
language en-US;
mode voice;
tag-format <lumenvox/1.0>;

root $FullTagGrammar;

$everyone = everyone:"everyone";
$having = having:"having";
$great = great:"great";
$time = time:"time";
$holidays = holidays:"holidays";
$later = later:"later";
$beach = beach:"beach";
$house = house:"house";
$barbeque = barbeque:"barbeque";
$beer = beer|beers:"beer";
$see = see:"see";
$soon = soon:"soon";

$WordsList = ($everyone|$having|$great|$time|$holidays|$later|$beach|$house|$barbeque|$beer|$see|$soon){$ =$$};


$Tag = [$WordsList{$.score = $$$.score}] $WordsList {$.wordslist = $$};

$FullTagGrammar = {$.tagslist =new Array} ($Tag {$.tagslist.push($$)} )<1-> [and $Tag{$.tagslist.push($$)}];



Sample Output
===================================================
 <tagslist><item index=0><score>[Undefined]</score><wordslist>having</wordslist></item><item index=1><score>[Undefined]</score><wordslist>time</wordslist></item></tagslist>


Back to top
Display posts from previous:   
   AsteriskGuru Archives Forum Index -> Asterisk-Speech-Rec All times are GMT
Page 1 of 1

 
Jump to:  
You can post new topics in this forum
You can reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2005 phpBB Group
contact us at: support@asteriskguru.com - asterisKGuru.com © all rights reserved   |   *asterisk is registered trademark of © Digium™