![]() |
|
![]() |
![]() |
|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Subscribe]
DM: Megaputer ships PolyAnalyst 4.0From: Sergei Ananyan Date: Tue, 18 Jan 2000 15:10:13 -0000
BW1594 JAN 18,2000 9:10 PACIFIC 12:10 EASTERN
( BW)(IN-MEGAPUTER-INTELLIGENCE) Megaputer Announces
a Final Release of PolyAnalyst 4.0 - the Leading Data Mining Solution
Business Editors & High Tech Writers
BLOOMINGTON, Ind.--(BUSINESS WIRE)--Jan. 18, 2000--
Megaputer announces the final release of PolyAnalyst 4.0, the newest
version of the leading data mining solution. An evaluation copy of
PolyAnalyst 4.0 can be downloaded from www.megaputer.com
Version 4.0 represents a major upgrade of the system, positioning
PolyAnalyst as the most comprehensive and versatile suite of data
mining algorithms available today. PolyAnalyst now utilizes Distributed
Component Object Model (DCOM) technology, features ten unique
machine learning algorithms, and furnishes versatile data manipulation,
visualization, and scoring capabilities. In addition to clustering,
predicting, dependency detecting, and yes/no classifying,
PolyAnalyst 4.0 solves tasks of explicit modeling, detection of
association rules in transactional data, and classification to multiple
categories. An open DCOM architecture makes PolyAnalyst 4.0
easily extendable, upgradable and customizable. It provides the
user with an option to purchase only the necessary machine
learning algorithms as individual modules and utilize these modules
as an integral part of their data storage and management system.
A DCOM-based PolyAnalyst Knowledge Server can support several
client stations on a local network.
New features of PolyAnalyst 4.0 include:
-- Interactive 3D graphics providing a better intuitive feeling for the
data prior to running machine learning analysis.
-- Unique Market Basket Analysis algorithm for processing transactional
data. Groups of products sold together well and the corresponding
directed association rules are identified - an order of magnitude faster
than by traditional algorithms.
-- Memory Based Reasoning algorithm based on a combination of Nearest
Neighbor and Genetic Algorithms. The new method is used efficiently for
classification into multiple categories, as well as prediction of
numerical
variables.
-- Implementation of DCOM architecture. Individual PolyAnalyst algorithms
can be easily integrated in the form of COM modules in external decision
support or data management applications. New PolyAnalyst machine
learning modules can be easily added and upgraded.
-- Support for OLE DB for Data Mining, a coming standard from Microsoft,
dramatically simplifying and unifying communications between RDBMS
and Data Mining applications.
-- Support for the analysis of large datasets. The maximum volume of data
accepted by the system has been significantly increased and a new
mechanisms for dealing with large datasets have been implemented.
-- Redesigned user interface. The project contents are organized in a
Windows-standard tree-like style, while preserving all the best
features of the traditional PolyAnalyst interface.
-- Dynamic HTML reports for exploration engines. The new interactive
reports can be customized, saved, printed, or copied in a standard
portable format, and exchanged with external applications
-- Mouse-driven Rule Assistant for simple creation of user defined
transformation rules.
-- Lift and Gain Charts for interpreting the results obtained by machine
learning algorithms are especially valuable for direct marketing tasks.
These features are very useful for measuring the extra profit reaped by
a marketer making decisions based on the knowledge discovered by
PolyAnalyst.
-- Snake Chart for convenient visual comparison of different data sets.
-- Chart Designer for the development of custom charts and advanced
data visualization capabilities
-- PA Scheduler enabling batch process data mining in PolyAnalyst.
The user can record a sequence of actions and schedule the created
script to be run by PolyAnalyst at a specified time and on the specified
datasets, applying the model to the specified data.
-- Sampling capabilities, allowing random selection of records from a
dataset.
-- Direct data exchange with Oracle Express and IBM Visual Warehouse.
"By offering these new features, Megaputer has clearly positioned
PolyAnalyst 4.0 as the number one modern business solution for data
analysis," says Sergei Arseniev, CEO of Megaputer Intelligence. "Now
PolyAnalyst combines the broadest selection of versatile machine
learning algorithms with the convenience and flexibility of DCOM
architecture. The new cutting-edge technology helped Megaputer
significantly increase its DM market share during the last year."
A number of Fortune 100 companies, such as Allstate Insurance, Boeing,
and DuPont Dow Elastomers, have already switched to PolyAnalyst 4.0.
Our customers are very enthusiastic about the system:
"Analytical engines do an excellent job of finding relations amongst
many fields without overfitting."
-- Timothy E. Nagle, Nycor Group, Consulting Scientist to 3M
"The software provides a unique and powerful set of tools for data
mining
applications, including promotion response analysis, customer segmentation
and profiling, and cross-selling analysis."
-- Raymond Burke, Chair of BA, Kelley School of Business, Indiana
University
Platforms: Windows NT/95/98/2000
Megaputer provides a complete family of innovative solutions for
Knowledge
Discovery in Databases, Data Mining, Semantic Text Analysis, and Information
Management. The offered solutions help users make better business decisions
by revealing knowledge hidden in their database or textbase. Megaputer
customers range from Fortune 100 companies, to numerous smaller businesses,
to government offices, and to educational institutions.
--30--slb/clv*
CONTACT: Megaputer Intelligence
Sergei Ananyan, 812/325-3026
FAX: 812/339-1646
s.ananyan@megaputer.com
www.megaputer.com
KEYWORD: INDIANA
INDUSTRY KEYWORD: COMPUTERS/ELECTRONICS SOFTWARE HARDWARE
TELECOMMUNICATIONS
|
MHonArc 2.2.0