William J. E. Potts

Learn More
There are several practical difficulties with the widespread application of artificial neural networks to predictive data mining. The inscrutability of the fitted model limits their suitability for many database marketing applications and can even have legal ramifications in credit scoring applications. Other difficulties concern determining the(More)
We present a visual tablet for exploring the nature of a bagged decision tree (Breiman [1996]). Aggregating classifiers over bootstrap datasets (bagging) can result in greatly improved prediction accuracy. Bagging is motivated as a variance reduction technique, but it is considered a black box with respect to interpretation. Current research seekine: to(More)
  • 1