Software
LaMa
I wrote the R package LaMa
for flexible model building and fast estimation of latent Markov models. It makes building such models much more convenient (kinda feels like stacking lego blocks) and allows for significant speed gains, so give it a try!
- Check out the package website here
- or get started with the vignette Intro to LaMa
- The package is motivated by the unification of various latent Markovian models developed in this paper
RTMBdist
I created the R package RTMBdist
, a library of probability distributions compatible with automatic differentiation delivered by the RTMB
package. While RTMB
allows flexible, fast likelihood-based modelling in R, many existing distributions break automatic differentiation. RTMBdist
fills this gap, enabling reliable and efficient model development.
EgoCor
I assisted Julia Dyck in developing the R package EgoCor
as a user-friendly interface based on the R package gstat
to fit exponential parametric models to empirical semi-variograms to model the spatial correlation structure of health data.