roams provides functions for estimating parameters and detecting outliers in state-space models using the ROAMS (Robust Outlier-Adjusted Mean-Shift) framework. It includes functionality for fitting benchmark models, visualizing model selection criteria, and evaluating in-sample and out-of-sample performance. Simulation tools for generating synthetic data under a first-differenced correlated random walk (DCRW) model are also included. Designed with flexibility for user-defined model structures via the specify_SSM() interface.
See the corresponding ROAMS paper published in Statistics and Computing here: https://doi.org/10.1007/s11222-026-10935-4
Installation
You can install the development version of roams from GitHub with:
# install.packages("pak")
pak::pak("rajan-shankar/roams")OR if using Windows:
# install.packages("devtools")
devtools::install_github("rajan-shankar/roams")