fontaine618/MSTweedie: Multi-source Sparse Tweedie Modelling

Extending the HDtweedie package to multi-source data, this package implements multi-source sparse Tweedie modelling. Given non-negative data with excess zeroes (typically insurance claim data), the algorithm fits the Tweedie distribution while penalizing the coefficient. Flexible regularization is achieved by an adjustable penalty factor that can balance between-feature sparsity and between-sources sparsity within features.

Getting started

Package details

AuthorSimon Fontaine, Yi Yang, Bo Fan, Wei Qian and Yuwen Gu.
MaintainerSimon Fontaine <[email protected]>
Package repositoryView on GitHub
Installation Install the latest version of this package by entering the following in R:
fontaine618/MSTweedie documentation built on April 4, 2018, 10:17 p.m.