Prev
Three-Day Data Mining Software Training Course:
The Salford Predictive Modeler Software Suite
Sponsored by:
Salford Systems
Dates: December 5-7, 2012
Location: Salford Systems' main office, San Diego, CA
Register: www.salford-systems.com/en/training/registration
December 5: Data Mining with Decision Trees (CART®)
Discover the power of tree structured data mining during this popular intro tutorial developed by Dan Steinberg, one of the world's leading experts in CART (classification and regression tree) technology and real world applications. This tutorial is geared toward statisticians and IT 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 decisions.
December 6: Predictive Modeling with MARS® Automated Non-linear Regression and RandomForests®
MARS
What is MARS? Why does it work? How can it be used? How can it help you develop more accurate regression models for problems such as predicting credit card holder balances, insurance claim losses, customer catalog orders, and cell phone use?
RandomForests
RandomForests®, created by Leo Breiman and Adele Cutler, is based on learning ensembles of CART trees. By judiciously injecting randomness into the tree-building process and then combining hundreds of these trees, RF is able to deliver high performance predictive models and a variety of novel exploratory data analysis results. RF also incorporates new metric free CLUSTER analyses that automatically select the variables used to define each cluster, with potentially different variables defining each cluster.
December 7: Advances in Data Mining (TreeNet®, GPS, Rulefit and Interaction Detection)
Attendees will be introduced to the main concepts in boosting methods in data mining. They will also be presented with the core innovations behind TreeNet stochastic gradient boosting, including the concepts of slow learning, use of weak learners in every stage of model building, resampling from the training at every stage, and ignoring data considered too far from the decision boundary in classification problems.