starvars: An R Package for Analysing Nonlinearities in Multivariate Time Series

Although linear autoregressive models are useful to practitioners in different fields, often a nonlinear specification would be more appropriate time series analysis. In general, there many alternative approaches nonlinearity modelling, one consists assuming multiple regimes. Among the possible specifications that account for regime changes multivariate framework, smooth transition most since they nest both and threshold models. This paper introduces [starvars](https://CRAN.R-project.org/package=starvars) package which estimates predicts Vector Logistic Smooth Transition model very general setting also includes predetermined variables. comparison existing R packages, starvars offers estimation of by maximum likelihood least squares. The allows test detect presence common breaks. Furthermore, computes multi-step-ahead forecasts. Finally, an illustration with financial is provided show its usage.