MSTweedie

R package for modeling multi-source insurance data

By Simon Fontaine, Yi Yang, Wei Qian, Yuwen Gu, and Bo Fan in R Package GitHub

September 14, 2022

The MSTweedie package provides multi-source sparse Tweedie modeling of insurance claim severity data with flexible regularization to balance between-feature sparsity and between-sources sparsity within features. An efficient group proximal gradient descent algorithm is used for fast computation.

Histogram of Tweedie data
Posted on:
September 14, 2022
Length:
1 minute read, 37 words
Categories:
R Package GitHub
Tags:
R Penalized regression Tweedie distribution Insurance
See Also:
SALES
gcdnet
A Unified Approach to Sparse Tweedie Modeling of Multisource Insurance Claim Data