Christoffer Moesgaard Albertsen

MSc Statistics, PhD,
Researcher at DTU Aqua

Software

To help my research, I develop software products in R and C++. Some of these products reach a stage where they can be useful for others. Below is a list of the software projects I have created that have reached this stage.

Software products I have contributed to or that are not mature yet can be seen on my GitHub profile.

argosTrack

Fit Movement Models to Argos Data for Marine Animals

Languages:

 Source code
 Latest release

Description:
argosTrack is an R package for fitting individual animal movement state-space models to data. The package depends on the TMB package for calculating likelihoods through the Laplace approximation. It was originally built for a paper, but has evolved a lot since. The package handles several continuous and discrete time models and several data types — despite the name.

covafillr

Local Polynomial Regression of State Dependent Covariates in State-Space Models

Languages:  

 Source code
 Latest release

Description:
The covafillr package is a wrapper for the covafill C++ library. The covafill library implements local polynomial regression and a search tree approximation to local polynomial regression. The C++ library includes interfaces for TMB and JAGS. The R package includes functions to use the C++ library from R along with convenience functions to combine functionality with ggplot2, Rcpp, TMB, and JAGS.

GrAD

Graph Based Reverse-Mode Automatic Differentiation

Languages:

 Source code

Description:
GrAD is a lightweight C++ library for reverse mode automatic differentiation. GrAD supports all basic C++ math operations and a few extra functions from the R math library. The library is still at an early stage.

otoclass

Otolith Classification and Proportion Estimation

Languages:

 Source code

Description:
The R package otoclass provides functions for extracting otolith contours from images and analysing them. The package includes methods for converting contours to Elliptical Fourier descriptors. These are used for classifying otoliths to fish stocks, and estimating stock mixing proportions. The otoclass package implements the methods needed for all parts of otolith shape analysis.