GBM is one of the most popular machine learning algorithms used in data mining competitions. Most of us use by way of it's R implementation (http://cran.r-project.org/web/packages/gbm/index.html). However, the folks at 0xData have recently written a distributed version for their open source h2o platform, and in the process become quite the experts on the algorithm! We are excited to have them stopping by to share their knowledge.
Earl Hathaway, 0xdata resident datascientist, will present on what types of problems GBM is applicable, and review recent KDD & Kaggle wins that used it.
Cliff Click, 0xdata CTO, will present about implementation and design choices of a Distributed GBM.
6:30 - 7:00pm Social (food + beer served)
7:00 - 8:30pm Talks
8:30 - 9:00pm Social