(411) previous ~ index ~ next

---

To: TDT Distrib <tdt-distrib@ldc.upenn.edu>
From: Mark Przybocki <mark.przybocki@nist.gov>
Subject: [Fwd: Re: Adjudicated TDT results are available]
Date: Mon, 28 Oct 2002 13:50:18 -0500

---

We are looking into the error pointed out by Jian Zhang. I've removed
the errorful results from our FTP site while we resolve this issue.

-mark

-------- Original Message --------
Subject: Re: Adjudicated TDT results are available
Date: Mon, 28 Oct 2002 13:11:16 -0500
From: Jian Zhang <jian.zhang@cs.cmu.edu>
To: Mark Przybocki <mark.przybocki@nist.gov>
References: <3DBD7289.E984A6F6@nist.gov>

Mark,

I just noticed that something goes wrong with NED's evaluation. If you
look at
CMU, PARC & UMASS's results, you'll
find that the minimum COST is worse than the actually COST. This will
happen if
the decisions are made by
several different thresholds. However, CMU only used a single threshold
to make
the decisions about whether
a story is new or not.

So, probably you want to check the evaluation to see what is wrong.

Bests,

-- Jian Zhang

CMU

Mark Przybocki wrote:

> Folks,
>
> We have rescored the TDT results with the adjudicated files provided to
> us by the LDC. You may download the rescored results:
>
> ftp://jaguar.ncsl.nist.gov/tmp/tdt2002_adjudicated_results_10.28.2002.tar.gz
>
> -mark
> --
> # // Mark Przybocki Phone: 301/975-3347
> # // NIST, Bldg 225 Rm B247 FAX: 301/670-0939
> # // Gaithersburg, MD 20899
> -------------------------------------------------------------
> To unsubscribe from tdt-distrib, email majordomo@ldc.upenn.edu
> with "unsubscribe tdt-distrib" in the body of the message.
-------------------------------------------------------------
To unsubscribe from tdt-distrib, email majordomo@ldc.upenn.edu
with "unsubscribe tdt-distrib" in the body of the message.

---

(411) previous ~ index ~ next

Last updated Mon Nov 11 14:16:29 2002