(086) previous ~ index ~ next

---

To: James Allan <allan@cs.umass.edu>
From: Jonathan Fiscus <jonathan.fiscus@nist.gov>
Subject: Re: TDT2 data confusion
Date: Wed, 15 Jul 1998 07:49:54 -0400

---

James,

James Allan wrote:
>
> Jon,
>
> > Most of the issues that you have raised will be solved by the updated
> > release of the scoring software and index files that I'm desparately
> > trying to get finished. I've finished the re-write of the segmentation
> > and tracking modules and I'm about to begin the detection module.
>
> This is great news. Thanks!
>
> I would expect a new release to resolve my first question ("when is a
> new release coming out"). The other two, though, ("what is manual
> transcript" and "specifying CCAP v. FDCH") seem like they also require
> some modification to the evaluation specification--they are not
> directly addressed in V2.7 of the document. If they'll be addressed
> in the new release, too, then we'll be happy to wait patiently....

We made an arbitrary decision to use the ccap transcripts as the source
of
choice for the eval. (The justification being the data will be "more"
similar
to the rest of the TV data.) However, we could make indexes available
that
use fdch transcripts as a contrast condition.

Jon



--
Jon Fiscus
NIST
Email: jfiscus@nist.gov
Phone: (301) 975-3182

---

(086) previous ~ index ~ next

Last updated Wed Sep 9 09:40:52 1998