(text)
Evaldas Verselis, Comdex Enterprise,
Frankfurt, Germany, Sep 28- Oct 1, 1998 http://www.comdex.de
--
on the latest news, publications, tools, meetings, and other relevant items
in the Data Mining and Knowledge Discovery field.
KD Nuggets is currently reaching over 5000 readers in 70+ countries
2-3 times a month.
Items relevant to data mining and knowledge discovery are welcome
and should be emailed to gps
in ASCII text or HTML format.
An item should have a subject line which clearly describes
what is it about to KDNuggets readers.
Please keep calls for papers and meeting announcements
short (50 lines or less of up to 80-characters), and provide a web site for
details, such as papers submission guidelines.
All items may be edited for size.
Back issues of KD Nuggets, a catalog of data mining tools
('Siftware'), pointers to data mining companies, relevant websites,
meetings, etc are available at KDNuggets Directory at http://www.kdnuggets.com/
********************* Official disclaimer ***************************
All opinions expressed herein are those of the contributors and not
necessarily of their respective employers (or of KD Nuggets)
*********************************************************************
~~~~~~~~~~~~ Quotable Quote ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
According to Reuters, Hong Kong bettors lost $52 million US dollars in
illegal bets on the World cup final when France beat favourites Brazil 3-0.
So much for safe bets in predicting the future ...
Previous1NextTop
Date: Sun, 12 Jul 1998 10:25:28 -0400
From: (Ismail Parsa)
Subject: KDD-CUP Last Call for Participation
Web: www.kdnuggets.com/meetings/kdd98/kdd-cup-98.html
| Sponsored by the |
| American Association for Artificial Intelligence (AAAI) |
| Epsilon Data Mining Laboratory |
| Paralyzed Veterans of America (PVA) |
+--------------------------------------------------------------------+
+--------------------------------------------------------------------+
| KDD-CUP Process and Important Dates |
+--------------------------------------------------------------------+
o Registration and signing of the NDA (Non-Disclosure Agreement)
July 1-15, 1998
We'll continue to accept registrations following the July 15 deadline but the
participants should be aware that they will have less time to analyze
the data as the results return date is August 14th.
o Release of the datasets (learning and validation), related
documentation and the KDD-CUP questionnaire
July 16, 1998
o Return of the results and the KDD-CUP questionnaire
August 14, 1998
o KDD-CUP Committee evaluation of the results
August 15-25
o Individual performance evaluations send to the participants
August 25, 1998
o Public announcement of the winners and awards presentation during
KDD-98 in New York City
August 29, 1998
My name is Frank Saeuberlich, I am working at the Institute for decision
theory and management science at the University of Karlsruhe.
To get a better insight into the field of KDD and Data Mining I am searching
for case studies and projects from this area.
In literature I found plenty examples which are mostly not very detailed.
Therefore it would be very helpful if someone could give me some references
or descriptions of practical projects by the following characteristics:
- Application area
- Problem description
- Data description (# records, attributes,...)
- Preprocessing methods
- Data Mining technique(s)/algorithm(s) used
- Postprocessing methods
- Evaluation criteria
- Software used
- Reference
Thank you very much.
Best regards from Karlsruhe,
Frank
Institut fuer Entscheidungstheorie und Unternehmensforschung
Universitaet Karlsruhe (TH)
Kollegium am Schloss, Bau III
Postfach 6980
D-76128 Karlsruhe
Telefon (Phone): +49/+721/+608-4284
Fax: +49/+721/+608-7765
e-mail: saeuber@etu.wiwi.uni-karlsruhe.de
In a joint pilot project between Silicon Graphics' and Risk Monitors
Inc, we looked at Loan Prepayment Modeling.
Loan Prepayment Modeling
Afshin Goodarzi(1) Ron Kohavi (2)
Richard Harmon (1) Aydin Senkut(2)
(1) Risk Monitors, Inc.
(2) Data Mining and Visualization, Silicon Graphics, Inc.
Loan level modeling of prepayment is an important aspect of hedging,
risk assessment, and retention efforts of the hundreds of companies in
the US that trade and initiate Mortgage Backed Securities (MBS). In
the pilot between Silicon Graphics Inc. and Risk Monitors Inc., we
used MineSet(TM) to look at different aspects of modeling customers who
have taken jumbo loans in the US. Using tools such as MineSet's Map
Visualizer, we show how refinancing costs differ across states and
counties. Using tools such as the Evidence Visualizer, we saw which
attributes make good predictor variables for prepayment forecasts. The
data comes from the McDASH Analytics database containing real
data, which tracks loans at monthly intervals.
The data in this file contains over 900,000 records by 45 attributes.
Here I like to mention about the publication of two new books on
feature transformation, published by Kluwer Academic Publishers that
can help people familiar with the on going research on data mining
especially on feature selection and extraction.
Feature Selection for Knowledge Discovery and Data Mining is intended to
be used by researchers in machine learning, data mining, knowledge
discovery and databases as a toolbox of relevant tools that help in
solving large real-world problems. This book is also intended to serve as
a reference book or secondary text for courses on machine learning, data
mining,
and databases.
The book can be used by researchers and graduate students in machine
learning, data mining, and knowledge discovery, who wish to understand
techniques of feature extraction, construction and selection for data
pre-processing and to solve large size, real-world problems. The book can
also serve as a reference work for those who are conducting research into
feature extraction, construction and selection, and are ready to meet the
exciting challenges ahead of us.
Presenting ANGOSS Software Corporation's KnowledgeSTUDIO, the next
generation of data mining tools:
* a workbench tool using decision trees, k-means clustering and neural networks
* support for application development and embedding within a large number of
programming languages
* rule generation/programming statements exported to Java, SAS, SQL and more
* ActiveX, IE4 and Netscape 3.0 enabled
* optional client/server interface and custom in-place mining drivers
KnowledgeSTUDIO's ease of use is enhanced by its graphical interface and its
use of wizards, making analysis a step by step process. KnowledgeSTUDIO's
advanced chart capabilities make data visualization a key advantage to the
company's understanding of the patterns within the information.
For further information check the web site at
-------------------------------------------------------
First educational version release of the DBMiner system
-------------------------------------------------------
DBMiner Technology Inc. is pleased to announce the release of the first
educational version of the DBMiner system, DBMiner E1.0. This version
consists of the following major modules: (1) data warehouse construction
(for automatic dimension generation and data cube creation), (2) 3-D
cube view of the data warehouse, (3) 3-D boxplot (statistical) view of
the data warehouse, (4) OLAP-based data summarizer, (5) associator (for
mining association rules), (6) classifier (for data classification and
decision tree construction), and (7) predictor (for regression analysis
and predictive modeling). Filtering, drilling, slicing, and dicing can
be performed on any part of the data cube during the OLAP and data mining
analysis. Clustering analysis, time-series analysis and statistical
analysis packages will be included in the future release.
The minimum system requirement for DBMiner is a Pentium-166 with 64 MB
RAM. The system runs on Windows/NT and Windows-95. DBMiner can be
directly linked to Microsoft SQLServer Version 6.5, or communicate with
various relational database systems, including Microsoft Access and
others, via ODBC connections.
A mini-version of DBMiner E1.0 (with the table size limited to 1000
rows and cube size limited to 3 dimensions) can be freely downloaded
at http://db.cs.sfu.ca/DBMiner
or http://www.dbminer.com.
A scalable
version of DBMiner E1.0 (Educational) can be purchased for educational
and research use at the price of $999.00 (U.S.) plus tax (when applicable),
which also covers one-year (up to July 1, 1999) free upgrade. We will
also take contracts for customized development of particular data mining
applications based on the DBMiner system.
This message is to announce the availablity of software written in
Fortran with an Splus interface implementing EM for parameterized Gaussian
mixture models + Poisson noise. A updated version of MCLUST for
model-based Gaussian hierarchical clustering is also included.
Functions for cluster analysis are also provided that combine hierarchial
clustering and EM using BIC (Bayesian Information Criterion) to determine the
model and number of clusters.
Chris Fraley (fraley@stat.washington.edu)
Statistics Dept., U. of Washington, Box 354322, Seattle, WA 98195-4322
Previous8NextTop
Date: Mon, 06 Jul 1998 12:01:13 -0400
From: Aviva Lev-Ari Aviva.Lev-Ari@Time-0.com
Subject: Employment Opportunity for an MSc or Ph.D. level
Web:
Industry has strond demand for applied recent graduates in
quantitative disciplines
I would like to explore the possibility of interviewing few of the
graduate students in the Stat/Math/OR/CS/Econometrics department.
Respectively, please ask the secretary of the Career Placement Center
to post an Ad, or e-mail to all graduate students in the above
departments, the following Job description.
Employment opportunity for a Stat/Math/OR/CS/Econ MSc or Ph.D. level.
Applied Research in Internet Economics and Electronic Commerce
Transaction Information Analytics and Data Mining
Profile:
Extremely bright, creative and inquisitive young broadly trained in
Quantitative Methods and Measurement Theory with an undergrade
education in Stat/Math/OR/CS/Econometrics/Psychometrics.
Modifyable into an independent applied researcher and heavy user of
S-Plus, SAS, Mathematica, MatLab, LaTex and graphical software.
Excellent writing (technical editorial skills) and oral communication
skills (ability to explain technical terms to non-technical
professionals). Independent in exploration of newly research concepts
assigned to, offer creative ideas to the project, and amenable to be
mentored and expand his/hers knowledge boundaries on a daily basis.
A team player, substantiated professional confidence, highest
integrity with handling data, choosing methods and respecting the
technical savvy of other peers and management.
Compensation:
Master Level: up to $45K - $60K
Ph.D. Level: up to $60K - $80K
Contact:
Aviva Lev-Ari, Ph.D.
Director of Information Analytics
Perot Systems Corp
101 Main St.
Cambridge, MA 02142
(617) 303-5011
UNIVERSITY OF PLYMOUTH
School of Electronic, Communication and Electrical Engineering
Research Assistant / Fellow in Virtual Data Mining
Job ref: 2816/TECH
Salary stlg10,018 to stlg15,411 pa - RA/RF
Required in the School of Electronic, Communication and Electrical
Engineering. The primary aim of the work is to investigate the
feasibility of developing a VDMT. You will join a team of researchers
who are the forerunners in establishing the field of Virtual Data
Mining.
The initial phase of the project will be 18 months, starting salary
will depend on experience.
You will have knowledge of data processing/analysis techniques.
A knowledge of C++ and OO would be an advantage.
Application Form and Further Particulars obtainable from the
Personnel Dept at the above address. Tel: 01752 232168, E-mail: personnel@plymouth.ac.uk
Please quote Ref and Job Title
The School of Information Systems, University of East Anglia, Norwich, has a
vacancy for a Research Assistant (Teaching Company Associate) to work on a
project entitled 'Datamining in the Telecommunications Sector'.
A computer graduate with at least a 2(I) degree in computing or allied
subject is sought for a two year post starting October 1st, 1998, or as soon
as possible thereafter.
The successful applicant will be an employee of the University of East
Anglia but will work within a leading telecommunications company, Nortel
plc, on a day-to-day basis. He/she will work at the Nortel site in Harlow,
in Essex, but will have regular visits to, and spend some time working with,
the Datamining Research Group at the University of East Anglia.
Opportunities will exist for registration for a part-time higher degree at
the University. A successful applicant will be expected to have a high
degree of numeracy and a strong computing background. Preference will be
given to those who, in addition, have some knowledge (and expertise) in one
or more of the following: optimisation, evolutionary computation, operations
research, artificial intelligence or telecommunications.
The research is sponsored jointly by the Teaching Company Scheme and by
Nortel plc and involves the development and application of various inference
and heuristic techniques, including genetic algorithms, simulated annealing
and tabu search, to elicit knowledge from large scale data sets generated
within the telecommunications industry.
Initial salary will be determined by experience but is expected to be in the
range 15,735 - 16,655 GBP, based on the Research and Analogous 1B scale.
Applicants are invited to telephone Dr George D Smith (01603 593260) or
email gds@sys.uea.ac.uk
for further information. Further information can
also be obtained from the www site:
Applications in the form of a covering letter plus three copies of a CV,
including the names and addresses of three referees, should be sent to:
Dr George D Smith
School of Information Systems
University of East Anglia
Norwich
NR4 7TJ
on or before Monday 20th July 1998. It is expected that interviews will take
place in the first two weeks of August and applicants are invited to
indicate in their letter of application their availability during this period.
The research group may have further opportunities for employment and
studentships in the area of data mining; details of the group can be found at: http://www.sys.uea.ac.uk/kdd
Enquiries concerning these opportunities may be made to Prof. Vic
Rayward-Smith (vjrs@sys.uea.ac.uk)
or Dr. George Smith.
----------------------------------------------------------------------------
Due to popular demand we announce a re-run of the seminar
ANALYSIS OF ENVIRONMENTAL DATA WITH MACHINE LEARNING METHODS
7.-10. September 1998, Ljubljana, Slovenia
----------------------------------------------------------------------------
Organized by Jozef Stefan Institute, Ljubljana,
in cooperation with University of Ljubljana
----------------------------------------------------------------------------
Area and goals of the seminar
The seminar will give an introduction to selected machine learning methods
as well as illustrative case studies of using these methods to analyse
environmental data, such as modeling algal growth in lakes and lagoons,
analysing the influence of physical and chemical parameters on selected
bioindicator organisms, and predicting the biodegradability of chemical
compounds. The participants will learn to use selected machine learning
tools and will have the opportunity for practical work with these tools on
real environmental data. The machine learning methods and tools introduced are
applicable to data analysis problems from different areas.
Who should attend the seminar
The seminar is intended for researchers and other professionals in the areas
of biology, chemistry, environmental science, and other areas related to
ecology and environmental management, whose work requires the analysis of
environmental data or modeling ecological processes.
What: M98 - Data Mining Technology Conference
When: October 26-30
Where: SAS Institute Inc., Cary, North Carolina
The conference features:
Herb Edelstein, President or Two Crows Corporation
Kurt Thearling, Director of Advanced Analytics at Exchange Applications
With best-practice presentations by industry leaders like:
Prudential Insurance Company of America
Bank of America
You should attend the conference if you want to:
* learn the role data warehousing plays in successful data mining
* learn how other organizations are using data mining effectively
* exchange ideas with industry experts and analysts
* identify key value factors in data mining technology.
Topics include:
* customer relationship management
* data warehousing and data mining: how they work together
* successful data mining strategies conducted in the real world.
Conference Options
October 26-30: Training Package - $1,500
(Conference plus both training courses)
October 27-30: Conference - $895
(Data Mining Technology Conference)
October 26: Pre-conference Training (1/2 day) - $150
Course: Data Mining Primer: Overview of
Applications and Methods)
October 29-30: Post-conference Training - $595
(Course: Using SAS Enterprise Miner: Applying
Data Mining Techniques)
Call for Papers: PADD99
The Third International Conference and Exhibition on
The Practical Application of Knowledge Discovery and Data Mining
Wednesday 21st April - Friday 23 April 1999
Commonwealth Insititute and Events Centre, London, UK http://www.demon.co.uk/ar/PADD99/
PADD99 is sponsored and supported to date by:
CompulogNet, LPA, WhiteCross Systems
PADD is a key part of the Practical Application Expo: a unique five-day
multi-technology, multi-track event which takes place every Spring in
London. PA Expo99 will also include conferences on Agents, Logic
Programming, Knowledge Management and JAVA.
You are invited to submit a paper to PADD99. Your paper should describe one
of the following:
* Commercially available products
* Internally deployed solutions
* Fully advanced pre-production prototypes
There are two types of submission available:
1.Paper
2.Industrial Report
Your submission can be on virtually any area or issue related to Knowledge
Discovery and Data Mining, which might include but are not limited to:
Telecommunications
Health Care
Marketing
Internet Based Mining
Finance
Retail
Customer Retention
Fraud Detection
Process Engineering/Industrial Control
Transport
In addition this year, we are expanding the conference to address business
issues in Data Mining.
Papers are therefore requested on Best Practices for Data Mining:
Managing Data Mining Projects
Cost Benefit Analysis
Project Success Factors
Risk Management
Achieving High Level Sponsorship
Knowledge Management
Integrating OLAP Solutions
Reasons for failure of Data Mining Projects
Especially of interest are papers that investigate the interface
between Knowledge Management, KDD and OLAP.
For further information and submission details please visit our web site at
the address above.
The Practical Application Company
PO Box 137
Blackpool
Lancs FY2 9UN
UK
Tel: +44 (0)1253 358081
Fax: +44 (0)1253 353811
email: info@pap.com
WWW: http://www.demon.co.uk/ar/TPAC/
The full conference program of COMDEX
Enterprise Frankfurt '98 to be held Sept. 28- Oct.1, 1998 is now
available on line at http://www.comdex.de
The program includes 7 conferences:
- COMDEX Internet
- Object World
- BankersIT Forum
- TelecomIT Forum
- META Briefing
- CA Enterprise Solutions
- Windows NT Forum
Evaldas Verselis
on behalf of Roberto Zicari
Chairman Advisory Board
COMDEX Enterprise Frankfurt
Knowledge Discovery Nuggets Index