Hi Curt,

Not would, we run collaboration network analyses for Biogen Idec researchers looking at a disease area or target.  We've even looked at treatment technologies to figure out who to work with or consult based on whether they are a supernode or provide significant linking potential.  We use Cytoscape for visualization and Medline or Dialog for providing the literature sources for analysis.  Our main area of improvement is normalizing the authors and affiliations more effectively.  It's not a major problem, but it would enhance the utility of the results.

It's similar in nature to what we've seen from Boston Consulting Group, just more flexible and internally produced.

William



On 7/21/06, Curt A. Monash <curtmonash@...> wrote:

At 04:33 AM 7/18/2006, William Hayes wrote:

Hi Mary,

Thank you for your introduction.  I noticed with significant interest your work on collapsing name variants in a systematic fashion.  One of our outstanding problems for generating collaboration networks using Cytoscape is just that issue of name variants (and making sure the names refer to specific individuals).  We expect that network analysis will play a part in this normalization.  I'd like to suggest Medline (the National Library of Medicine's biomedical abstracts database) containing author and affiliation information with no associated unique primary ID's as an excellent database to perform this work :)

If you do find this interesting as part of your graduate studies, I'd be happy to help you get started.

Regards,

William

---

 

That sounds pretty interesting William.  What kind of network analysis would you do?

Thanks,

CAM

Curt A. Monash, Ph.D.
President, Monash Information Services
curtmonash@...
(978) 266-1815 (main)
Backups:  curtmonash@..., (978) 266-1866

Blogs: 
http://www.monashreport.com (Software industry, IT-related public policy, ventures, etc.)
http://www.dbms2.com (DBMS and related technologies)
http://www.texttechnologies.com (Search, text mining, etc.)
http://www.softwarememories.com (Software industry history)

Computerworld columns: http://www.computerworld.com/action/columnist.do?command=viewColumnist&bylineID=894
Core website (slightly out of date) http://www.monash.com

Hi Mary,

Thank you for your introduction.  I noticed with significant interest your work on collapsing name variants in a systematic fashion.  One of our outstanding problems for generating collaboration networks using Cytoscape is just that issue of name variants (and making sure the names refer to specific individuals).  We expect that network analysis will play a part in this normalization.  I'd like to suggest Medline (the National Library of Medicine's biomedical abstracts database) containing author and affiliation information with no associated unique primary ID's as an excellent database to perform this work :)

If you do find this interesting as part of your graduate studies, I'd be happy to help you get started.

Regards,

William

---

 

Hi all,

Just a short introduction of myself and a response to Seth's question.
I'm a senior lecturer at Macquarie University, Sydney, Australia,
where I am doing research in the area of question answering.

And my answer to Seth's question is, besides suggesting you to look at
my own question answering project
<http://www.ics.mq.edu.au/~diego/answerfinder/>, to look at the TREC
question answering track of the TREC conferences
<http://trec.nist.gov/>, or at the CLEF conferences
<http://clef.isti.cnr.it/>, where much of current research in QA is
published.

Text-based question answering is a very dynamic area of R&D, and the
main web search engines are starting to incorporate QA technology.
Expect to see more and more QA abilities in the future web search
engines. Current systems focus on short fact-based questions, and
currently they are starting to attempt more complex questions where
the answer needs to be composed from bits and pieces found in various
sources.

The following list of QA systems is not exhaustive but it can give you
an idea of what you can find nowadays.

http://www.ics.mq.edu.au/~pizzato/repository

Hope this helps.

Cheers,

Diego

--- In TextAnalytics@yahoogroups.com, Seth Grimes <grimes@...> wrote:
>
> Hello all,
>
>  What's going on in the world of natural-language query/question
> answering?  For examples of this, see (and try!) --
>
> http://start.csail.mit.edu/
>
> http://brainboost.com
>
> http://answers.com
>
>  For that matter, go to http://google.com and enter "2 + 2 - 1/17"
> or "map Georgia."  Google Enterprise has partners that are extending
this
> capability to cover artifacts produced in response to a Google OneBox
> "search."  If I understand this correctly, the partner's software
inserts
> Google index entries and Google lists those artifacts along with
document
> hits.
>
>  I'm interested in particular in implementations for governmental /
> social / economic statistics (and maps).  The sites I cited do
alright for
> simple questions about demographics but fail on more complex but still
> typical questions.  I'd guess that's because they're broadly targeted;
> perhaps they could be tuned for the vocabularies and syntaxes of stats
> questions.
>
>  I'd like to hear about academic and industrial research and
> productizations and to get pointers to papers.
>
>  Thanks,
>
> 				 Seth
>
>
> --
> Seth Grimes   Alta Plana Corp, analytical computing & data management
>               Intelligent Enterprise magazine (CMP), Contributing Editor
> grimes@...       http://altaplana.com    301-270-0795
>

I know also about Ngram Statistics Package (NSP).

"The Ngram Statistics Package (NSP) is a suite of programs that aids in
analyzing Ngrams in text files. We define an Ngram as a sequence of 'n'
tokens that occur within a window of at least 'n' tokens in the text;
what constitutes a "token" can be defined by the user.'

Project page:
http://search.cpan.org/~tpederse/Text-NSP-0.97/Docs/README.pod#DESCRIPTION

roxana

> I'm working on a catalog of open-source software for text-analytics and
> related functions.  Here's what I have so far.  Please add to the list.
> After a bit more review, I'll paste this info into the Wikipedia Text
> Analytics entry.
>
> It would be great to have your reactions to the various packages!
>
> 				 Seth
>
>
> OpenNLP
>
> "An organizational center for open source projects related to natural
> language processing....  OpenNLP also hosts a variety of java-based NLP
> tools which perform sentence detection, tokenization, pos-tagging,
> chunking and parsing, named-entity detection, and coreference using the
> OpenNLP Maxent machine learning package."
>
> Home page: http://opennlp.sourceforge.net/
>
> Project page: http://sourceforge.net/projects/opennlp/
>
>
> Carrot2
>
> "A search results clustering framework. Includes clustering components and
> a stand-alone meta search component. Combines well with indexing and
> search engines (open source and proprietary)."
>
> Home page: http://www.carrot2.org
>
> Project page: http://sourceforge.net/projects/carrot2/
>
>
> FreeLing
>
> "An open source language analysis tool suite."
>
> http://garraf.epsevg.upc.es/freeling/
>
>
> GATE -- General Architecture for Text Engineering
>
> "GATE is ... the leading toolkit for Text Mining ... comprised of an
> architecture, a free open source framework (or SDK) and graphical
> development environment."
>
> Home page: http://gate.ac.uk/index.html
>
> Project page: http://sourceforge.net/projects/gate
>
>
> Graphviz -- Graph Visualization Software
>
> Graph visualization is a way of representing structural information as
> diagrams of abstract graphs and networks....  The Graphviz layout programs
> take descriptions of graphs in a simple text language, and make diagrams
> in several useful formats such as images and SVG for web pages, Postscript
> for inclusion in PDF or other documents; or display in an interactive
> graph browser."
>
> http://graphviz.org
>
>
> jTokeniser
>
> "The jTokeniser package was designed to combine a set of tokenisers that
> range from basic whitespace tokenisers to more complex ones that deal
> intuitively with natural language....  Tokenisers include:
>
> * WhiteSpaceTokeniser
> * StringTokeniser (based on specified delimiters)
> * RegexTokeniser (regular expression defines a token)
> * RegexSeparatorTokeniser (define what is *not* a token)
> * BreatIteratorTokeniser (sophisticated locale-specific tokeniser)
> * SentenceTokeniser (sentence segmentation)"
>
> http://www.andy-roberts.net/software/jTokeniser/
>
>
> Kea
>
> "Kea-3.0 automatically extracts keyphrases from the full text of
> documents....  Kea-4.0 is a new version of Kea that has been developed for
> controlled indexing of documents in the domain of agriculture."
>
> http://www.nzdl.org/Kea/
>
>
> LingPipe ** free but not open source
>
> "A suite of Java libraries for the linguistic analysis of human language."
>
> http://www.alias-i.com/lingpipe/index.html
>
>
> LTC -- Linguistic Tree Constructor
>
> "LTC is a free program for building linguistic syntax trees from text."
>
> Home page: http://ltc.sourceforge.net
>
> Project page: http://sourceforge.net/projects/ltc
>
>
> Lucene
>
> "Apache Lucene is a high-performance, full-featured text search engine
> library written entirely in Java."
>
> http://lucene.apache.org/
>
>
> NLTK
>
> "NLTK, the Natural Language Toolkit, is a suite of program modules, data
> sets and tutorials supporting research and teaching in computational
> linguistics and natural language processing."
>
> Home page: http://nltk.sourceforge.net/index.html
>
> Project page: http://sourceforge.net/projects/nltk
>
>
> Nutch
>
> "Nutch builds on Lucene Java to provide web search application software."
>
> http://lucene.apache.org/nutch/
>
>
> TouchGraph
>
> "TouchGraph provides a hands-on way to visualize networks of interrelated
> information. Networks are rendered as interactive graphs, which lend
> themselves to a variety of transformations."
>
> Home page: http://www.touchgraph.com/
>
> Project page: http://touchgraph.sourceforge.net/
>
>
> Weka
>
> "Weka is a collection of machine learning algorithms for data mining
> tasks.... Weka contains tools for data pre-processing, classification,
> regression, clustering, association rules, and visualization. It is also
> well-suited for developing new machine learning schemes."
>
> http://www.cs.waikato.ac.nz/~ml/weka/
>
> See Weka-related projects:
> http://weka.sourceforge.net/wiki/index.php/Related_Projects
>
>
>
> --
> Seth Grimes   Alta Plana Corp, analytical computing & data management
>               Intelligent Enterprise magazine (CMP), Contributing Editor
> grimes@...       http://altaplana.com    301-270-0795
>
>
>
>
> Yahoo! Groups Links
>
>
>
>
>
>
>

I forgot to include UIMA --

UIMA -- Unstructured Information Management Architecture

"An open, industrial-strength, scaleable and extensible platform for
creating, integrating and deploying unstructured information management
solutions from combinations of semantic analysis and search components."

Project site at IBM Research:  http://www.research.ibm.com/UIMA/

SDK site at IBM alphaWorks:  http://www.alphaworks.ibm.com/tech/uima

Framework site at SourceForge:  http://uima-framework.sourceforge.net





On Wed, 19 Jul 2006, Seth Grimes wrote:

> I'm working on a catalog of open-source software for text-analytics and
> related functions.  Here's what I have so far.  Please add to the list.
> After a bit more review, I'll paste this info into the Wikipedia Text
> Analytics entry.
>
> It would be great to have your reactions to the various packages!
>
> 				 Seth
>
>
> OpenNLP
>
> "An organizational center for open source projects related to natural
> language processing....  OpenNLP also hosts a variety of java-based NLP
> tools which perform sentence detection, tokenization, pos-tagging,
> chunking and parsing, named-entity detection, and coreference using the
> OpenNLP Maxent machine learning package."
>
> Home page: http://opennlp.sourceforge.net/
>
> Project page: http://sourceforge.net/projects/opennlp/
>
>
> Carrot2
>
> "A search results clustering framework. Includes clustering components and
> a stand-alone meta search component. Combines well with indexing and
> search engines (open source and proprietary)."
>
> Home page: http://www.carrot2.org
>
> Project page: http://sourceforge.net/projects/carrot2/
>
>
> FreeLing
>
> "An open source language analysis tool suite."
>
> http://garraf.epsevg.upc.es/freeling/
>
>
> GATE -- General Architecture for Text Engineering
>
> "GATE is ... the leading toolkit for Text Mining ... comprised of an
> architecture, a free open source framework (or SDK) and graphical
> development environment."
>
> Home page: http://gate.ac.uk/index.html
>
> Project page: http://sourceforge.net/projects/gate
>
>
> Graphviz -- Graph Visualization Software
>
> Graph visualization is a way of representing structural information as
> diagrams of abstract graphs and networks....  The Graphviz layout programs
> take descriptions of graphs in a simple text language, and make diagrams
> in several useful formats such as images and SVG for web pages, Postscript
> for inclusion in PDF or other documents; or display in an interactive
> graph browser."
>
> http://graphviz.org
>
>
> jTokeniser
>
> "The jTokeniser package was designed to combine a set of tokenisers that
> range from basic whitespace tokenisers to more complex ones that deal
> intuitively with natural language....  Tokenisers include:
>
> * WhiteSpaceTokeniser
> * StringTokeniser (based on specified delimiters)
> * RegexTokeniser (regular expression defines a token)
> * RegexSeparatorTokeniser (define what is *not* a token)
> * BreatIteratorTokeniser (sophisticated locale-specific tokeniser)
> * SentenceTokeniser (sentence segmentation)"
>
> http://www.andy-roberts.net/software/jTokeniser/
>
>
> Kea
>
> "Kea-3.0 automatically extracts keyphrases from the full text of
> documents....  Kea-4.0 is a new version of Kea that has been developed for
> controlled indexing of documents in the domain of agriculture."
>
> http://www.nzdl.org/Kea/
>
>
> LingPipe ** free but not open source
>
> "A suite of Java libraries for the linguistic analysis of human language."
>
> http://www.alias-i.com/lingpipe/index.html
>
>
> LTC -- Linguistic Tree Constructor
>
> "LTC is a free program for building linguistic syntax trees from text."
>
> Home page: http://ltc.sourceforge.net
>
> Project page: http://sourceforge.net/projects/ltc
>
>
> Lucene
>
> "Apache Lucene is a high-performance, full-featured text search engine
> library written entirely in Java."
>
> http://lucene.apache.org/
>
>
> NLTK
>
> "NLTK, the Natural Language Toolkit, is a suite of program modules, data
> sets and tutorials supporting research and teaching in computational
> linguistics and natural language processing."
>
> Home page: http://nltk.sourceforge.net/index.html
>
> Project page: http://sourceforge.net/projects/nltk
>
>
> Nutch
>
> "Nutch builds on Lucene Java to provide web search application software."
>
> http://lucene.apache.org/nutch/
>
>
> TouchGraph
>
> "TouchGraph provides a hands-on way to visualize networks of interrelated
> information. Networks are rendered as interactive graphs, which lend
> themselves to a variety of transformations."
>
> Home page: http://www.touchgraph.com/
>
> Project page: http://touchgraph.sourceforge.net/
>
>
> Weka
>
> "Weka is a collection of machine learning algorithms for data mining
> tasks.... Weka contains tools for data pre-processing, classification,
> regression, clustering, association rules, and visualization. It is also
> well-suited for developing new machine learning schemes."
>
> http://www.cs.waikato.ac.nz/~ml/weka/
>
> See Weka-related projects:
> http://weka.sourceforge.net/wiki/index.php/Related_Projects
>
>
>
> --
> Seth Grimes   Alta Plana Corp, analytical computing & data management
>               Intelligent Enterprise magazine (CMP), Contributing Editor
> grimes@...       http://altaplana.com    301-270-0795
>
>
>
>
> Yahoo! Groups Links
>
>
>
>
>
>
>

--
Seth Grimes   Alta Plana Corp, analytical computing & data management
               Intelligent Enterprise magazine (CMP), Contributing Editor
grimes@...       http://altaplana.com    301-270-0795

I'm working on a catalog of open-source software for text-analytics and
related functions.  Here's what I have so far.  Please add to the list.
After a bit more review, I'll paste this info into the Wikipedia Text
Analytics entry.

It would be great to have your reactions to the various packages!

					 Seth


OpenNLP

"An organizational center for open source projects related to natural
language processing....  OpenNLP also hosts a variety of java-based NLP
tools which perform sentence detection, tokenization, pos-tagging,
chunking and parsing, named-entity detection, and coreference using the
OpenNLP Maxent machine learning package."

Home page: http://opennlp.sourceforge.net/

Project page: http://sourceforge.net/projects/opennlp/


Carrot2

"A search results clustering framework. Includes clustering components and
a stand-alone meta search component. Combines well with indexing and
search engines (open source and proprietary)."

Home page: http://www.carrot2.org

Project page: http://sourceforge.net/projects/carrot2/


FreeLing

"An open source language analysis tool suite."

http://garraf.epsevg.upc.es/freeling/


GATE -- General Architecture for Text Engineering

"GATE is ... the leading toolkit for Text Mining ... comprised of an
architecture, a free open source framework (or SDK) and graphical
development environment."

Home page: http://gate.ac.uk/index.html

Project page: http://sourceforge.net/projects/gate


Graphviz -- Graph Visualization Software

Graph visualization is a way of representing structural information as
diagrams of abstract graphs and networks....  The Graphviz layout programs
take descriptions of graphs in a simple text language, and make diagrams
in several useful formats such as images and SVG for web pages, Postscript
for inclusion in PDF or other documents; or display in an interactive
graph browser."

http://graphviz.org


jTokeniser

"The jTokeniser package was designed to combine a set of tokenisers that
range from basic whitespace tokenisers to more complex ones that deal
intuitively with natural language....  Tokenisers include:

* WhiteSpaceTokeniser
* StringTokeniser (based on specified delimiters)
* RegexTokeniser (regular expression defines a token)
* RegexSeparatorTokeniser (define what is *not* a token)
* BreatIteratorTokeniser (sophisticated locale-specific tokeniser)
* SentenceTokeniser (sentence segmentation)"

http://www.andy-roberts.net/software/jTokeniser/


Kea

"Kea-3.0 automatically extracts keyphrases from the full text of
documents....  Kea-4.0 is a new version of Kea that has been developed for
controlled indexing of documents in the domain of agriculture."

http://www.nzdl.org/Kea/


LingPipe ** free but not open source

"A suite of Java libraries for the linguistic analysis of human language."

http://www.alias-i.com/lingpipe/index.html


LTC -- Linguistic Tree Constructor

"LTC is a free program for building linguistic syntax trees from text."

Home page: http://ltc.sourceforge.net

Project page: http://sourceforge.net/projects/ltc


Lucene

"Apache Lucene is a high-performance, full-featured text search engine
library written entirely in Java."

http://lucene.apache.org/


NLTK

"NLTK, the Natural Language Toolkit, is a suite of program modules, data
sets and tutorials supporting research and teaching in computational
linguistics and natural language processing."

Home page: http://nltk.sourceforge.net/index.html

Project page: http://sourceforge.net/projects/nltk


Nutch

"Nutch builds on Lucene Java to provide web search application software."

http://lucene.apache.org/nutch/


TouchGraph

"TouchGraph provides a hands-on way to visualize networks of interrelated
information. Networks are rendered as interactive graphs, which lend
themselves to a variety of transformations."

Home page: http://www.touchgraph.com/

Project page: http://touchgraph.sourceforge.net/


Weka

"Weka is a collection of machine learning algorithms for data mining
tasks.... Weka contains tools for data pre-processing, classification,
regression, clustering, association rules, and visualization. It is also
well-suited for developing new machine learning schemes."

http://www.cs.waikato.ac.nz/~ml/weka/

See Weka-related projects:
http://weka.sourceforge.net/wiki/index.php/Related_Projects



--
Seth Grimes   Alta Plana Corp, analytical computing & data management
               Intelligent Enterprise magazine (CMP), Contributing Editor
grimes@...       http://altaplana.com    301-270-0795

William,

Thank you for your reply.

I'm sure that at some point my work will expand to include structured
data like citations, which have similar problems no matter what the
domain.  But for now my work is limited to unstructured full text
documents in the domain of genealogy, so MedLine is not in the cards for
my dissertation, but perhaps for the future beyond my dissertation.

My husband is on the faculty of SUNY Upstate Medical University here in
Syracuse, so I imagine at some point I will get involved with medical or
perhaps biomedical domain work.  During dinner with his colleagues one
night I was trying to explain my dissertation topic, and one of his
fellow researchers immediately drew a parallel with genes and proteins.

If you'd like, I can point you to literature that does involve work with
citations -- there's quite a bit out there.

-- Thanks again,
     Mary D. Taffet
     Ph.D. Candidate/Syracuse University School of Information Studies
     Scientist/TextWise LLC
     Syracuse, NY


William Hayes wrote:
> Hi Mary,
>
> Thank you for your introduction.  I noticed with significant interest
> your work on collapsing name variants in a systematic fashion.  One of
> our outstanding problems for generating collaboration networks using
> Cytoscape is just that issue of name variants (and making sure the names
> refer to specific individuals).  We expect that network analysis will
> play a part in this normalization.  I'd like to suggest Medline (the
> National Library of Medicine's biomedical abstracts database) containing
> author and affiliation information with no associated unique primary
> ID's as an excellent database to perform this work :)
>
> If you do find this interesting as part of your graduate studies, I'd be
> happy to help you get started.
>
> Regards,
>
> William
>
>
> [snip]

Hello,

My name is Mary D. Taffet.  I have a Bachelor's degree in Linguistics
from UNC-Chapel Hill, a Master's degree in Linguistics from Syracuse
University, an MLS degree in Information and Library Science from
Syracuse University's School of Information Studies and am currently a
Ph.D. Candidate at the School of Information Studies.

In between my bachelor's and master's programs, I became a business
applications programmer working with COBOL on a mainframe.  Needless to
say, at some point I realized that I didn't have to choose between
working with language and working with computers, both of which I both
enjoy and am fairly good at.  So I went back to school with the goal of
learning about Natural Language Processing/Computational Linguistics,
which I have been focusing on since 1999.  [And in the process became
one of the few skilled COBOL programmers to never ever work on a Y2K
project, though I got offers most every week it seemed...]

I was a Research Assistant at TextWise from 1999-2000, then was a
Research Assistant at the Center for Natural Language Processing at
Syracuse University's School of Information Studies from 2000-2004.  Now
I'm back at TextWise as a fulltime employee since 2005 working on
contextual advertising.

At some point after I started grad school, I became addicted to
genealogy, and have done the bulk of my genealogical research online
since then, with a few trips to Salt Lake City and the Montreal Archives
along the way.  I am a very frustrated online genealogical researcher
due to the difficulty of searching names online.  Fortunately the
difficulty in searching names online is something that even a
non-genealogical researcher can do as a dissertation as it is a general
problem for all sorts of applications.  So that's the focus of my
dissertation.  I am looking at the relationship between people and the
way people are referred to in written documents.  I hope to bring
together all variant forms of a person's name, while at the same time
teasing apart identical names that refer to different people.  I have an
electronic corpus of 14,000+ biographies from a 1904 publication
supplied by Ancestry.com.

I had to put my dissertation work aside for a while during my father's
hospitalization last year, and am still trying to get back into the
swing of things after my father passed away.  It's not easy with a
fulltime job, but hopefully I will get there before too much more time
has passed.

-- Mary D. Taffet
     Ph.D. Candidate/Syracuse University-School of Information Studies
     Scientist/TextWise LLC
     Syracuse, NY

---------- Forwarded message ----------
Date: Mon, 17 Jul 2006 16:37:12 -0500
From: "Marsh, Brice" <Brice.F.Marsh@...>

  I'm academicaly curious, but I don't have the time to become an active
participant. My name is Brice Marsh and I'm the Executive Director of Teen
Think Tanks of America, Inc. (www.teenthinktanks.org) and we generate lots
of collaborative material that we need to classify and organize for
reporting purposes. However, my "day job" is as a senior computer
scientist with a federal contractor for NASA at Marshall Space Flight
Center and that keeps me busy. But, I do wish to be able to stay abreast
of your work and the progress of your research. So in this regard, I must
be classified as a "taker" and not a "giver", I'm sorry; but you're more
than welcome to review/use any of the material we have posted on the TTT
website, only with attribution, please.

Thanks.

Brice F. Marsh
bricemarsh@...

Tam,

Text Tiling is obviously a good choice. However, as far as I know, the
implementation of this method is not easy, the algorithm is
time-consuming, and the results can be unpredictible.

Have you thought about simply dividing your documents into paragraphs
or (overlapping or non-overlapping) window passages (i.e. sequences of
words)?

Regards,
Dominic

--- In TextAnalytics@yahoogroups.com, tamer adel <tamadel2003@...> wrote:
>
> Hi,All
>   I have text document of one mass and  i want to divide it into
multi paragaph that are coherent portions....the subject is new to me
....i made search and i found text tiling algorithm is preferred
method to execute my task ... is any one know more than me guid  to
another method  or algorithm as i suggested to solve the suggestion
problem.
>   pls, replay is urgent to me till the afternoon of tomorrow.
>
>   regards,
>   tam
>
>
>  Tamer Abu Elenain
>       Software Developer
>       (+2) 012 562 74 21
>    tamadel2003@...
>
>
>
> ---------------------------------
> Yahoo! Messenger with Voice. Make PC-to-Phone Calls to the US (and
30+ countries) for 2¢/min or less.
>

Hi William,
You might want to check into Schemalogic. They are based in Kirkland
WA. www.schemalogic.com

Regards,
Lew Larson
--- In TextAnalytics@yahoogroups.com, "William Hayes"
<william.s.hayes@...> wrote:
>
> Hi all,
>
> Has anyone run across a good thesauri  or ontology management server.
> Something that will allow user editing of hierarchically tagged
canonical
> names with their synonyms, basic visualization and can export into
various
> formats such as ANSI Thesaurus format and is accessible via web
services for
> accessing synonym suggestions for enterprise search engines?  I know
that's
> a pretty tall order, but it's one that we need in the text analytics
area.
> I've run across some fairly expensive tools for this, but I was
hoping to
> find something a good deal less pricey (and hopefully easily
extendable).
>
> We need to manage protein, disease, tissue, cell line, adverse event,
> pathological process, etc thesauri.  I'd like to be able to tag the
protein
> thesauri with various relationship information such as Pathway
> participation, Molecular function, Biological process, etc with is-a and
> part-of relations maintained and be able to tag particular synonyms with
> filterable labels (such as 'rarely used', 'ambiguousWithGeneralText',
> 'ambiguousWithOtherProtein','stopword', etc.).  This would really be a
> terminological resource server for text analytic engines.  The more
easily
> edited (in an ad hoc) fashion, the easier it can be set up as a public
> resource to allow (moderated?) community curation of terminologies
for text
> mining.
>
> TIA,
>
> William
>

Hello to Karl Wiig, Seth Grimes, Neil Raden, and everyone I've not yet had the
pleasure of meeting either electronically or in person,

I'm Joe Firestone, Managing Director and CEO of Center for the Open Enterprise,
LLC. COE is the parent company of the Knowledge Management Consortium
International (KMCI), the KMCI Publishing Group (KMCI Press and KMCI Online
Press) and the new Adaptive Metrics Center. Both KMCI and AMC do independent
research and also offer training and consulting services. KMCI in Knowledge
Management and AMC in Business Performance Management and Measurement.

My interest in text analytics goes back to the 1950s when I first learned about
content analysis applied to Soviet studies. Later on, I did some research for
the US Air Force on Intentions Analysis and Forecasting, and on applying
computerized content analysis to the study of national intentions and long-range
forecasting of inter-nation behavior. In the early to middle 70s, I published a
few academic articles using measures of national motives in statistical models
predicting national behavior. Since then, my changing interests have carried me
into many other areas, but I've always tried to keep up with the progress of
text analytics.

More recently, my work in Knowledge Management and Adaptive Metrics has led me
back to a greater focus on text analytics, since I'm persuaded that if you want
to measure the quality of problem and knowledge claim formulation, and also the
quality of knowledge claim evaluation, one of the best ways to develop metrics
is through analysis of the semantic patterns in text. This idea has an important
place in our training workshops and in our treatment of core software tools for
knowledge management in KMCI's CKIM Certificate Training Workshop. It is also
the idea I'll be pursuing most often in this group.

Best,


Joe

Joseph M. Firestone, Ph.D.
Managing Director, CEO
KMCI and the Adaptive Metrics Center
www.kmci.org
www.adaptivemetricscenter.com
http://radio.weblogs.com/0135950

CKO
Executive Information Systems, Inc.
www.dkms.com
703-461-8823

Please visit our web site:   WWW.CONSULTBIZ.COM

"Performance Improvement through Technology Planning and Operational Redesign"

* Certified Management Consultant (CMC) is a certification mark awarded by the Institute of Management Consultants USA and represents evidence of the highest standards of consulting, and his adherence to the technical and ethical canons of the profession.  Less than 1% of all management consultants have achieved this level of performance. Certified Computing Professional (CCP) is awarded by the Institute for the Certification of Computing Professionals, and certifies proficiency in the information technology field.

Hello all,

	 What's going on in the world of natural-language query/question
answering?  For examples of this, see (and try!) --

http://start.csail.mit.edu/

http://brainboost.com

http://answers.com

	 For that matter, go to http://google.com and enter "2 + 2 - 1/17"
or "map Georgia."  Google Enterprise has partners that are extending this
capability to cover artifacts produced in response to a Google OneBox
"search."  If I understand this correctly, the partner's software inserts
Google index entries and Google lists those artifacts along with document
hits.

	 I'm interested in particular in implementations for governmental /
social / economic statistics (and maps).  The sites I cited do alright for
simple questions about demographics but fail on more complex but still
typical questions.  I'd guess that's because they're broadly targeted;
perhaps they could be tuned for the vocabularies and syntaxes of stats
questions.

	 I'd like to hear about academic and industrial research and
productizations and to get pointers to papers.

	 Thanks,

					 Seth


--
Seth Grimes   Alta Plana Corp, analytical computing & data management
               Intelligent Enterprise magazine (CMP), Contributing Editor
grimes@...       http://altaplana.com    301-270-0795

Why not use a Java-based solution so that it just doesn't matter?

Hi all,

I am new to this group. Here are a few words about myself:

I hold a Ph.D. in cognitive computer science from Université du Québec à
Montréal. In my doctoral dissertation ("Application de techniques de forage
de textes de nature prédictive et exploratoire à des fins de gestion et
d'analyse thématique de documents textuels non structurés"), I explored and
validated the use of descriptive and predictive text mining techniques to
assist thematic analysis of unstructured documents. My current research
interests concern the use of hybrid text mining techniques (using concepts
and techniques from both linguistics and artificial intelligence) to assist
ontology development from unstructured documents. I also collaborate in
various projects concerning the application of text mining techniques in the
context of institutional repositories and digital libraries,
computer-assisted reading and text analysis, etc.
I am currently a postdoctoral fellow at Observatoire de Linguistique
Sens-Texte (Université de Montréal) and will be (starting December 1st,
2006) assistant professor at École de Bibliothéconomie et des Sciences de
l'Information (Université de Montréal).

Regards,
Dominic Forest

¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Dominic Forest
Chercheur postdoctoral
Observatoire de Linguistique Sens-Texte (OLST)
Université de Montréal
 
Courrier électronique : dominic.forest@...
Sites Internet : www.dominicforest.com
_____________________________________________________
 

Whereas we help clients with "Windows" and Linux environments, we use
Mac OS 10.4.7 exclusively.  Any suggestions for what you pursue?

Greetings
--
Karl M. Wiig
Chairman
Knowledge Research Institute, Inc.
7101 Lake Powell Drive, Arlington, TX 76016 USA
Phone: (817) 572-6254 / Cell: (682) 554-3998 / Fax: (817) 478-1048
http://www.krii.com

A quick welcome to the Text Analytics discussion group.  I think you'll
find we have a nice mix of researchers, vendors, and practioners here;
also some recruiters.

While I set up the group and Rob Raisch, CTO of Compliance Week is serving
as co-moderator, I hope to do very little moderating.  You're free to post
whatever statements, questions, problems, and announcements you wish so
long as they relate to text analytics as you define that term.  Just
follow the usual rules regarding respect for other list members, and
please identify yourself when posting unless you have a good reason not
to.  Do introduce yourself to the list if you wish.

Thanks all,

				 Seth Grimes


--
Seth Grimes   Alta Plana Corp, analytical computing & data management
               Intelligent Enterprise magazine (CMP), Contributing Editor
grimes@...       http://altaplana.com    301-270-0795