Random Forests was developed by Leo Breiman, one of the original CART(tm) "gang of four," in the early 2000's. A commercial version is available from Salford Systems, who also commercialized CART (and are licensees of the TM's). However, Leo's original FORTRAN code is freely available and is the basis of a port to R by Andy Liaw and Matt Wiener of Merck in 2001. An independent implementation, cforest, is in the R package "party" by Hothron, Hornik & Zeileis in Vienna. Diaz-Uriarte has recently extended randomForest for variable selection purposes.
Random Forests compares favorably with other machine learning methods in predictive power with advantages of ease of setup, allowing continuous and categorical variables (both dependent & independent), natural resistance to over fitting, usable for both supervised and unsupervised learning, and visual interpretation of variable importance and partial variable dependence. The latter two can provide important business concepts about the model.
After a brief introduction to Loyalty Matrix and R, this virtual seminar will look at Random Forests in detail and walk through a number of case studies based on actual projects. We will demonstrate some actual runs in R so the participants will get a realistic view of using R in general and Random Forests in particular.
Presentation is informal and participant interaction is encouraged.
Speaker:
Jim Porzak, VP Analytics, Loyalty Matrix, Inc
Jim Porzak, VP Analytics at Loyalty Matrix, Inc. in San Francisco has been using R for customer analytics since 2000. He has presented at both of the useR! User Conferences to date. At this spring's conference he gave one of the pre-conference tutorials titled "Using R for Customer Analytics."
Ways to Register
| Online: | Register Now |
| Fax: | 212.302.7643 |
| Phone: | 212.790.1500 |
| Mail: | Printable Form DMA Customer Service 1120 Avenue of the Americas New York, NY 10036-6700 (Check payable to DMA) |
For questions/inquiries, call DMA Customer Service at 212.790.1500 or e-mail customerservice@the-dma.org.
Payment Policy and Confirmation
In order to confirm your place in this event, we require payment in full. From time to time, we change a date or location of an event. If we need to change a date or location for any reason, you will be contacted. If you have not received confirmation of your attendance from DMA, please call 212.790.1500 to ensure your seat.
Cancellation Policy
If you must cancel, please submit cancellations in writing to The DMA Customer Service Dept. Registrations canceled at least 5 days before the event will be refunded 100%. Cancellations received less than 5 days before the event will not be refunded, rather the remainder to be held on account for a future DMA event within 12 months.
To Transfer or Change Your Registration: Send a written request on company letterhead at least seven days before the event to DMA Customer Service via fax at 212.302.7643 or mail to: DMA Customer Service, 1120 Avenue of the Americas, New York, NY 10036-6700. Within seven days of the event, all changes will be made on-site.
DMA Money-Back Guarantee
You will be 100% satisfied with what you learn or we will refund your registration fee in full.* We can afford to make this offer because we know that this seminar will exceed your expectations. It's part of our commitment to providing you with the highest possible quality in education and training.
*Requires written request within 30 days of the seminar.
© Direct Marketing Association | Privacy Statement