Tutorial on

Data Mining Techniques

Part of the 2004 Advanced Simulation Technologies Conference (ASTC'04)

Sponsored by:
The Society for Modeling and Simulation International (SCS)

Sunday, April 18th, 2004, 7 pm to 9 pm
Monday, April 19th, 2004, 7 pm to 9 pm
Tuesday, April 20th, 2004, 7 pm to 9 pm

Hyatt Regency Crystal City
Arlington, Virginia

presented by
Mikhail Golovnya, Senior Statistician
Salford Systems

Data Mining supports simulation in very definitive ways. It is often far more powerful to integrate data mining with simulation than just using simulation alone. By using Data Mining, the professional can develop fused data-physical models and or models based solely on data sets or data streams. The derived models can then be used as either high-fidelity approximations to slower running models, models that describe the information content of the data set, or as a method / subroutine in a larger simulation model.

In this tutorial series you will learn how to use a variety of the latest Data Mining tools to improve upon your existing Simulation Models. You will learn how to quickly and easily discover patterns and relationships in your data that you never knew existed. You will learn how to create models in minutes that previously may have taken you days or weeks. You will also be given free access to 3 advanced Data Mining programs for 30 days. At the end of each of these seminars, you will have access to the software as well as the understanding to be able to try out these tools on your own data.

Sunday: Data Mining with Decision Trees and how to build and interpret CART models

Discover the power of tree-structured data mining during this popular introductory seminar, geared toward audiences who are interested in understanding the conceptual basis of decision tree technology -- what it is, why it works, how it has been used, and how it can help you make better business and/or research decisions. Explore the practical use and application of decision trees for solving real world data mining problems.

Monday: Predictive Modeling with Automated Non-linear Regression Tools and how to build and interpret MARS models

MARS automates the development and deployment of accurate and easy to understand regression models. Conventional regression models typically fit straight lines to data. Although this usually oversimplifies the data structure, the approximation is sometimes good enough for practical purposes. However, in the frequent situations in which a straight line is inappropriate, an expert modeler must search tediously for transformations to find the right curve. MARS can quickly trace out any pattern detected in the data. Topics covered include: What is MARS? Why does it work? How can it be used? How can it help you develop more accurate regression models?

Tuesday: Advanced Data Mining Techniques and how to build and interpret TreeNet/MART and Random Forests models.

Learn how to use Data Mining software recently developed at Stanford University and Berkeley by world-renowned statisticians Leo Breiman and Jerome Friedman. You will learn how to use TreeNet/MART and Random Forests. Topics also include: hybridizing data mining tools, understanding what advanced features to look for in data mining software and how to use these advanced features.

Cost: free for registered conference attendees.

Sponsored by The Society for Modeling and Simulation International
P.O. Box 17900
San Diego, California 92177
Phone 858-277-3888
Fax 858-277-3930
E-mail scs@scs.org

Return to ASTC'04 Home Page

Return to SCS Home Page

Mission Membership Organization Conferences Publications Resources

Copyright 2004, SCS. All Rights Reserved.