[ANN] LikelihoodProfiler.jl Major Upgrade

Package Announcements
, ,
1

Hi everyone,

We have released a new version of LikelihoodProfiler.jl, which represents a major redevelopment of the package.

In its initial (old) version, LikelihoodProfiler.jl provided a method (CICO) for estimating confidence intervals and assessing the practical identifiability of model parameters using the profile likelihood approach. In this new version, we have introduced various methods for computing likelihood profiles and estimating confidence intervals. These methods are now unified under a new interface, making it easier for users to switch between them.

This release marks the first version of the revamped package. We plan to introduce additional methods, improve interfaces, enhance documentation, and more. If this package is relevant to your work, we will be grateful for your feedback and suggestions for improvement. Your input will help us make the package more valuable to the community.

10 Likes
2

Cool! Any hope to have this interfaced with Turing.jl models?:eyes:

3

Hi,

Thank you for your interest in LikelihoodProfiler.jl! If I understand correctly, you’re looking to use profiling in the mode estimation part of Turing.jl, where a Maximum Likelihood Estimation (MLE) approach is used—is that right? Could you open an issue on the LikelihoodProfiler.jl with an example? We can work through it and identify what’s missing or needs to be added on the LikelihoodProfiler side.