Tweedie's Compound Poisson Model with Grouped Elastic Net

• Activity Number: 696
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistical Computing
• Author(s): Wei Qian* and Yi Yang and Hui Zou
• Companies: Rochester Institute of Technology and University of Minnesota and University of Minnesota
• Keywords: coordinate descent ; insurance score ; IRLS-BMD ; lasso ; risk segmentation ; variable selection

Abstract

Tweedie's Compound Poisson model is a popular method to model data with probability mass at zero and non-negative, highly right-skewed distribution. Motivated by wide applications of the Tweedie model in various fields such as actuarial science, we investigate the grouped elastic net method for the Tweedie model in the context of the generalized linear model. To efficiently compute the estimation coefficients, we devise a two-layer algorithm that embeds the blockwise majorization descent method into an iteratively re-weighted least square strategy. The proposed algorithm is implemented in an easy-to-use R package HDtweedie, and is shown to compute the whole solution path very efficiently. The modeling applications in risk segmentation of insurance business are illustrated by analysis of an auto insurance claim dataset.