File /Humanist.vol22.txt, message 213


Date:         Fri, 26 Sep 2008 07:39:37 +0100
From: Humanist Discussion Group <willard.mccarty-AT-MCCARTY.ORG.UK>
Subject: 22.223 text-analysis in the news, and new software
To: humanist-AT-Princeton.EDU


               Humanist Discussion Group, Vol. 22, No. 223.
       Centre for Computing in the Humanities, King's College London
                        www.princeton.edu/humanist/
                     Submit to: humanist-AT-princeton.edu

   [1]   From:    Humanist Discussion Group                           74)
                 <willard.mccarty-AT-mccarty.org.uk>
         Subject: Re: 22.209 text-analysis in the news

   [2]   From:    Humanist Discussion Group                           22)
                 <willard.mccarty-AT-mccarty.org.uk>
         Subject: JGAAP 3.1. now available!


--[1]------------------------------------------------------------------
         Date: Fri, 26 Sep 2008 07:05:44 +0100
         From: Humanist Discussion Group <willard.mccarty-AT-mccarty.org.uk>
         Subject: Re: 22.209 text-analysis in the news
         In-Reply-To: <48CAA335.3090207-AT-mccarty.org.uk>

[Apologies for the hiatus, which makes the following less easy to
follow. See the quoted text for context. --WM]

This is true, but some caveats are in order. Campaign rhetoric is a
highly ritualized genre that lends itself to easy statistical analysis
and visualization. If you want to be a little cynical, you might say
it's like those tests that the Pentagon has set up to prove that anti-
missile systems work-- not entirely real life experiments with the
thumbs weighing in heavily on the side of success.

Things are harder and a lot less obvious once you get into the messy
universe of literary texts in different genres, from different places,
and different times. We have been valiantly wrestling with this set of
problems in the MONK project (http://monkproject.org), and with a
little luck we hope to demonstrate real life success stories in the
few months.

I don't want to knock the Times visualizations, they are very good,
and they certainly are a sign of changing times.  But alas, from a
standpoint of text analysis,  any American political campaign is like
shooting fish in a barrel.


On Sep 12, 2008, at 12:13 PM, Humanist Discussion Group wrote:

>             Humanist Discussion Group, Vol. 22, No. 209.
>     Centre for Computing in the Humanities, King's College London
>                      www.princeton.edu/humanist/
>                   Submit to: humanist-AT-princeton.edu
>
>
> ), and
>       Date: Fri, 12 Sep 2008 18:04:09 +0100
>       From: Humanist Discussion Group <willard.mccarty-AT-mccarty.org.uk>
>       >
> The New York Times has a variant that illustrates beautifully the way
> that visual representation of complex data is becoming more
> common/accepted. The web teaches us to write as well as read.
> http://www.nytimes.com/interactive/2008/09/04/us/politics/20080905_WORDS_GRAPHIC.html
>
> Steve
>
> Stephen Woodruff
> Humanities Advanced Technology & Information Institute
> 11 University Gardens
> University of Glasgow
> Scotland/UK G12 8QQ
> +44 (0) 141 339 8855
> www.hatii.arts.gla.ac.uk
>
>> -----Original Message-----
>> From: Humanist Discussion Group [mailto:humanist-AT-Princeton.EDU] On =20
>> Behalf Of Humanist Discussion Group
>> Sent: 11 September 2008 16:17
>> To: humanist-AT-Princeton.EDU
>> >                Humanist Discussion Group, Vol. 22, No. 204.
>>      Centre for Computing in the Humanities, King's College London
>>                       www.princeton.edu/humanist/
>>                    Submit to: humanist-AT-princeton.edu
>>        Date: Thu, 11 Sep 2008 16:15:05 +0100
>>        From: Humanist Discussion Group <willard.mccarty-AT-mccarty.org.uk=20
>> >
>>        Subject: Text analysis
>> I notice the increasing use of simple text-analytic results in =20
>> journalism, probably just because it's gotten so easy to do.  Today =20
>> I heard about one that lets you compare McCain and Obama, though =20
>> it's not so good at showing you any of the context:
>> http://www.speechwars.com/
>> John
>> --
>> Dr John Lavagnino
>> Senior Lecturer in Humanities Computing
>> Centre for Computing in the Humanities, King's College London
>> 26=C3=83=C2=82=C3=A2=C2=80=C2=9329 Drury Lane
>> London WC2B 5RL
>> +44 20 7848 2453
>> www.lavagnino.org.uk
>> General Editor, The Oxford Middleton
>>    http://www.oup.com/uk/catalogue/?ci=3D9780198185697
>>    http://www.oup.com/uk/catalogue/?ci=3D9780198185703





--[2]------------------------------------------------------------------
         Date: Fri, 26 Sep 2008 07:08:07 +0100
         From: Humanist Discussion Group <willard.mccarty-AT-mccarty.org.uk>
         Subject: JGAAP 3.1. now available!
         In-Reply-To: <48CAA335.3090207-AT-mccarty.org.uk>

We are delighted to announce the release of a new version of JGAAP,
the _Java Graphical Authorship Attribution Program_, by Patrick
Juola and his research team at Duquesne University.  For more details,
please visit our wiki at www.jgaap.com, from which you can also download
both the software and the source code.

JGAAP, now in version 3.1, is a graphical program to perform text-based
authorship attribution among other forms of text categorization
using a variety of different methods.  It uses the Java programming
language to create an easily-extensible framework for solving text
classification problems in a write-once, run-anywhere fashion. Currently
available analytic methods include popular classification techniques
such as nearest-neighbor algorithms using a variety of distances,
support vector machines using a variety of kernels, with many others
planned for addition in the near future.

This program is freely available and released under Open Source guidelines;
we hope that other researchers will help us test and extend this software
for widespread use.

Thank you,


Patrick Juola and the JGAAP development team
(John Noecker, Mike Ryan, Chuck Liddell, Sandy Speer, and Ashley Bernola)
www.jgaap.com

(JGAAP is supported by National Science Foundation award #OCI-0721667.)

From - Fri Sep 26 07:58:47 2008
X-Mozilla-Status: 0000
X-Mozilla-Status2: 00000000
   

Humanist Main Page

 

Display software: ArchTracker © Malgosia Askanas, 2000-2005