I tried it again copying the numbers many, many, many more times, threw in
more instances of my Social Security number, and a whole bunch of other just
random numbers, including numbers with decimal points. It seems to work.
No more "for" as the last digit!
I did not use the topic editor when I first installed Dragon. Since I got a
new computer at about the same time, and did not copy over all my old files,
I did not have a lot to work with. Now that I do have more files laying
around I ran it for the first time. It is doing the number thing better,
but I'm finding it making mistakes it didn't used to. It seems to be miss
recognizing several custom words I had added to the vocabulary. When you
run the topic builder, does it un-learn everything it had already learned?
It still seems to know the words are there, but they are no longer
"preferred". They now show up toward the bottom of the list in a correction
box, when they use to be the first choice.
Steve Hasselberg
SHasselb@@cldx.com
----------
From: Holland St.John
To: Hasselberg, Stephen
Cc: Voice Users
Subject: Re: Number Recognition
Date: Tuesday, March 24, 1998 9:24PM
Steve...
I'm not sure why you are getting "for" at the end of a series of
numbers. When I try that with my version I get the correct numeral 4
had the end of the series of dictated numbers.
It may be that I use a fair number of numerals in my dictation in
addition to the words for and four. If you rarely use numerals, perhaps
the language model you have generated gives preference to the word "for"
over the numeral "4".
You might be into try to fix this by generating a document that has the
numbers 0-9, then copy and paste it many times so they have a document
with lots of numbers in it. Then run the vocabulary builder on this to
document in addition to all your old documents. This will generate a
new language model that includes numerals. ie
0 1 2 3 4 5 6 7 8 9
0 1 2 3 4 5 6 7 8 9
0 1 2 3 4 5 6 7 8 9
0 1 2 3 4 5 6 7 8 9
0 1 2 3 4 5 6 7 8 9
0 1 2 3 4 5 6 7 8 9
etc.
Holland
Hasselberg, Stephen wrote:
>
> I'm having some trouble with recognizing numbers in NaturallySpeaking. My
> Social Security number, for example, ends in the number 4. It almost
always
> adds the word " for" to the end of the previous digits. Why does it do
> this? Doesn't it realize that I've been dictating a string of numbers and
> this is most likely a number too! I'm certain I'm not pausing before the
> last digit.
>
> Steve Hasselberg
> SHasselb@@cldx.com
------ Message Header Follows ------
Received: from smtp.cnp.cldx.com by msmail.cnp.cldx.com
(PostalUnion/SMTP(tm) v2.1.8d for Windows NT(tm))
id AA-1998Mar24.212751.1268.665262; Tue, 24 Mar 1998 21:27:51 -0500
Received: from fw.cldx.com (fw [151.41.2.4]) by smtp.cnp.cldx.com
(8.6.12/8.6.10) with ESMTP id VAA13596 for <SHasselb@@cldx.com>; Tue, 24 Mar
1998 21:17:48 -0500
Posted-Date: Tue, 24 Mar 1998 21:17:48 -0500
Received: from localhost by fw.cldx.com id VAA12032; Tue, 24 Mar 1998
21:08:15
-0500 sender holland@@rattlers.net for <SHasselb@cldx.com>
Received: from unknown(204.96.161.2) by fw.cldx.com via smap (g3.0.1)
id sma012022; Tue, 24 Mar 98 21:07:59 -0500
Received: from [204.96.161.230] by host.corridor.net (NTMail 3.02.13) with
ESMTP id da190921 for <SHasselb@@cldx.com>; Tue, 24 Mar 1998 20:24:10 -0600
Message-ID: <35186788.902@@rattlers.net>
Date: Tue, 24 Mar 1998 20:10:16 -0600
From: "Holland St.John" <holland@@rattlers.net>
Organization: San Marcos Diagnostics
X-Mailer: Mozilla 3.0 (Win95; I)
MIME-Version: 1.0
To: "Hasselberg, Stephen" <SHasselb@@cldx.com>
CC: Voice Users <voice-users@@voicerecognition.com>
Subject: Re: Number Recognition
References: <1998Mar24.132133.1268.1098287@@msmail.cnp.cldx.com>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
![]() |