Methods for M-estimation of statistical models
MEstimation is a Julia package that implements M-estimation for statistical models (see, e.g. Stefanski and Boos, 2002, for an accessible review) either by solving estimating equations or by maximizing inference objectives, like likelihoods and composite likelihoods (see, Varin et al, 2011, for a review), using user-specified templates of just
- the estimating function or the objective functions contributions
- a function to compute the number of independent contributions in a given data set
A key feature is the use of those templates along with forward mode automatic differentiation (as implemented in ForwardDiff) to provide methods for reduced-bias M-estimation (RBM-estimation; see, Kosmidis & Lunardon, 2020).
See the documentation for more information, and the examples for a showcase of the functionality MEstimation provides.
See NEWS.md for changes, bug fixes and enhancements.
| Ioannis Kosmidis | (author, maintainer) |
|---|---|
| Nicola Lunardon | (author) |
- Varin C, Reid N, and Firth D (2011). An overview of composite likelihood methods. Statistica Sinica 21(1), 5-42. Link
- Kosmidis I, Lunardon N (2020). Empirical bias-reducing adjustments to estimating functions. ArXiv:2001.03786. Link
- Stefanski L A and Boos D D (2002). The calculus of M-estimation. The American Statistician(56), 29-38. Link