Predictive Modeling in Databases

 

Byung S. Lee and Zhen He

 

Abstract

“Predictive modeling” in this presentation refers to building a statistical or mathematical model of a quantity with the objective of using the model to predict the quantity at a given future time.  It finds a full range of applications in databases and data mining. My current research activities center on this theme. This presentation will include overviews of a recent project on predictive modeling of database function costs (funded by DoE) and a new project on forecasting-based predictive modeling for several database problems (funded by NSF). The presentation will also give an outline of previous and present collaborations with other researchers.