U.S. Department of Energy

Pacific Northwest National Laboratory

Towards Automated Feature Engineering for Transactional Data

Monday, April 6, 2015
Rob Jasper
Advances in machine learning have led to some of the most recognized and profitable systems in computing. Despite these advances, feature engineering has largely remained a manual, but critical step in development of these systems. This talk presents an automated method for constructing features from transactional data based on parameterized feature templates. Templates can represent a near infinite number of features compactly. Previous experiments have shown this approach to be highly competitive. A framework for generalizing this approach is proposed.
Speaker Bio

Rob Jasper is Vice President of Data Sciences at compensation analytics provider PayScale. Previous to joining PayScale, he served as Chief Technology Officer of Intelligent Results which produced one of the first web-based statistical modeling and decision support solutions. Intelligent Results was acquired by First Data in 2007. At First Data, Rob served as Vice President of Engineering for Information Services and CTO of the Analytics Center of Excellence where he oversaw engineering for more than 20 products in the areas of fraud, risk, analytics, and data services. Earlier in his career, Rob was an adjunct faculty member at Seattle University and researcher at The Boeing Company.

| Pacific Northwest National Laboratory