The Fast RODEO for Local Polynomial Regression

An open challenge in nonparametric regression is finding fast, computationally efficient approaches to estimating local bandwidths for large data sets, in particular in two or more dimensions. In the work presented here we introduce a novel local bandwidth estimation procedure for local polynomial regression which combines the greedy search of the RODEO algorithm with linear binning. The result is a fast, computationally efficient algorithm we refer to as the <i>fast RODEO</i>. We motivate the development of our algorithm by using a novel scale-space approach to derive the RODEO. We conclude with a toy example and real world example using data from the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observation (CALIPSO) satellite validation study, where we show the fast RODEO’s improvement in accuracy and computational speed over two other standard approaches.