R package: numerical.mle

A set of numeric MLE solvers.

This is very early alpha. I just started this project and it is not ready for use yet. I just took a bunch of numerical code from algebraic.mle and put it in this separate package. I will be adding more numerical solvers and more examples in the future. Most of the code probably does not even work yet, since I haven’t tested it.

Installation

You can install numerical.mle from GitHub with:

install.packages("devtools")
devtools::install_github("queelius/numerical.mle")
Copy

API

A set of methods for fitting log-likelihood functions to data. We provide various adapters for log-likelihood functions, including penalty adapters (for constrained MLEs) and transformation adapters (for transformed MLEs).

The object representing a fitted model is a type of mle object, the maximum likelihood estimator of the model with respect to observed data. We use the R package for this purpose. (See here).

The API mostly consists of generic methods with implementations for various mle type objects. For a full list of functions, see the function reference for numerical.mle.

Examples

Fitting a linear regression model

library(numerical.mle)
library(algebraic.mle)
Copy