Random forests

Supervisor contact

Related areas

Regression trees are piecewise-constant regression models relating the response to the predictors by partitioning the space of predictors and random forests are sets of trees grown on bootstrap samples from the data that together often give better accuracy than individual trees. Possible projects can explore the methodology and the use of trees, recent theoretical work on trees and recent work on the relationship between trees and neural networks.

This project will be co-supervised with Dr Francis Hui.

Updated:  23 June 2017/Responsible Officer:  Director/Page Contact:  School Manager