Activity Number:
|
588
- A Mixed Bag of Graphical Delights
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 1, 2018 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Statistical Graphics
|
Abstract #329875
|
Presentation
|
Title:
|
Plotting Two-Dimensional Confidence Regions
|
Author(s):
|
Christopher Weld* and Lawrence Leemis and Andrew Loh
|
Companies:
|
William & Mary and William & Mary and William & Mary
|
Keywords:
|
Graphical Methods;
Parameter Estimation;
Numerical Optimization
|
Abstract:
|
Plotting two-parameter confidence regions (CR) is non-trivial. Numerical methods rely on a computationally expensive grid-like exploration of the parameter space to develop its contours. A recent advance reduces the two-dimensional problem to a series of one-dimensional problems employing a trigonometric transformation that assigns an angle from the maximum likelihood estimator (MLE), and an unknown radial distance to its CR boundary. This paradigm shift improves plot accessibility by easing the computational burden by three orders of magnitude, but parameter scale differences make it susceptible to poor definition and/or computational inefficiencies given a naive approach to its chosen set of angles. This work improves the low cost radial profile log likelihood plot technique by selectively targeting points along the CR boundary. Two heuristics are given: an elliptic-inspired angle selection heuristic and an intelligent confidence region smoothing search heuristic. Each improves graphic quality and computation time over the established technique, and are automated in R and publicly available via the Comprehensive R Archive Network (CRAN) 'conf' package's crplot() function.
|
Authors who are presenting talks have a * after their name.