Colloquium Details
Post Doc Colloquium: Active Learning with Adaptive Heterogeneous Ensembles
| Author: | Zhenyu Lu University of Vermont |
|---|---|
| Date: | May 19, 2011 |
| Time: | 09:00 |
| Location: | 220 Deschutes |
| Host: | Dejing Dou |
Abstract
Classification techniques build predictive models with data described by a set of features (attributes) and associated labels (a discrete set of possible classes). One popular approach to classification is ensemble methods, which instead of relying on one single classification model such as Decision Trees (DT), combine a set of models for prediction. Ensemble methods have been successfully applied in many fields other than classification, such as clustering, relevance ranking and recommendation systems. An open question in ensemble methods is how to choose one model type (homogeneous ensemble), or a set of model types (heterogeneous ensemble) to construct ensembles. Our research addressed four fundamental questions for heterogeneous ensembles: 1) if we need heterogeneous ensembles: we proved that heterogeneous ensembles could outperform homogeneous ensembles of any involving classification model alone; and 2) how to construct appropriate heterogeneous ensembles: we introduced an algorithm called Adaptive Heterogeneous Ensembles (AHE) to automatically discover appropriate combinations of classification model types; and 3) why heterogeneous ensembles work: through analysis we concluded that heterogeneous ensembles outperform homogeneous ensembles because different classification model types complement each other; and 4) when heterogeneous work: we discovered that the advantage of heterogeneous ensembles over other methods is increased when the target data have more class labels. In our work, the efficacy of AHE is experimentally validated in the context of active learning.
Biography
Zhenyu Lu is a Ph.D. candidate under the direction of his advisors Professor Josh Bongard and Professor Xindong Wu at the University of Vermont. He expects to obtain his Ph.D. degree by August, 2011. His research interests are ensemble methods and active learning for supervised learning. His Ph.D. thesis studies how to effectively construct heterogeneous ensemble for active learning.