DM: RE: Datamining Definition.

We have been calling text mining a process that can transform text into a
form amenable to analysis, understanding, and prediction.  The goal is to
create an information infrastructure that can integrate the data extracted
from text and transactions with analysis or data mining tools.

One means of creating the data is through the use of  entity extraction.
Entity extraction is a natural language processing technique that
automatically recognizes the names of people, places, things, etc. from text
and then stores them in a database.  The text extraction software
automatically determines the concepts (such as people, places, things,
organizations, affiliations, etc.) that are embedded within the text and the
relationships between those concepts.  This type of search is especially
ideal for those instances where the user does not know a priori all of the
information that they would like to search for.

We have been using sophisticated fact extraction software developed by Liz
Liddy called KnowIt.  KnowIt handles the extraction of semantic information
(from a variety of standard news feeds) and its storage, for rapid
searching, in a relational database.  Software sub-components recast raw
text feeds into a common SGML-tagged format, identify complete sentences,
tag each sentence word with its part-of-speech, identify and interpret
proper nouns and several kinds of phrases, extract relational data from the
tagged text, and store this data using an Oracle database management system.
We use the Oracle database as the basis for data mining visualization.

Lisa Sokol
Technical Director
Information Analysis Solutions Sector
Veridian MRJ Technology Solutions
(703) 277-1888 Fax: (703) 277-1472

-----Original Message-----
From: owner-datamine-l@nautilus-sys.com
[mailto:owner-datamine-l@nautilus-sys.com]On Behalf Of Franklin Wayne
Poley
Sent: Tuesday, March 21, 2000 4:35 PM
To: datamine-l@nautilus-sys.com
Subject: DM: Datamining Definition.



Could datamining be fairly defined as "extracting data of value from
text"? (With "text" broadly defined to include words, sentences, pictures,
symbols etc?
FWP.

http://users.uniserve.com/~culturex/Machine-Psychology.htm

• Prev by Date: Re: DM: Datamining Definition.
• Next by Date: DM: CEC2000 - Challenge and Competitions
• Prev by thread: DM: CEC2000 - Challenge and Competitions
• Next by thread: Re: DM: Datamining Definition.
• Index(es):
  • Date
  • Thread