Software

NetVAR

This R-package is a software to model Vector AutoRegressive (VAR) models which are fit to data with a structure typical for networks. As the method contained in NetVAR is tailored to network structures, in the presence of such structures the model fit is improved as well as forecasting accuracy. This is particularly true in the presence of high-dimensional data and when only short time series are available. Further the model improves interpretability of the underlying data. The embedded LISAR method employs methods from the sparsity literature such as LASSO, SCAD and Adaptive LASSO. The package automatically derives the required regularisation sequences and refines it during the estimation to provide the optimal model. The package allows for model optimisation under various loss functions such as MSFE, AIC, and BIC. It provides a dedicated class, allowing for summary prints of the optimal model and a plotting function to conveniently analyse the optimal model via heatmaps.

The package is available via CRAN: https://github.com/SimonTrimborn/NetVAR  

gofCopula

This R-package is a fire-and-forget software for Goodness-of-Fit (GoF) tests for Copulae. It consists of 13 tests and a Hybrid test, which allows for an informed decision about the functional form of the Copula. The package features in-build parallelization of the bootstrapping tasks, seed control for full reproducibility of the results and allows for the integration of user-specified GoF tests. The latter allows for the creation of new GoF tests and a benchmarking study of the new test with the package. 

The package is available via CRAN: https://CRAN.R-project.org/package=gofCopula 

IndexConstruction

This R-package is an implementation of the method developed for the CRIX index, Trimborn and Härdle (2018). The package allows for the derivation of a financial index with a flexible number of constituents, such that the index represents the current state of the market well. The package allows for a full derivation of an index and has individual functions for updating an existing index. The latter comes in handy for the updating of an index on a website. The package was the underlying engine for the website https://thecrix.de/ . CRIX and VCRIX were bought in 2021 by Royalton Partners and are derived by S&P Global nowadays.

The package is available via CRAN: https://CRAN.R-project.org/package=IndexConstruction