Contributed Poster Presentations: Astrostatistics Interest Group

iAR is an R and Python package that provides tools for handling autoregressive models for one or two dimensional irregularly observed stationary time series. Note that the standard autoregressive (AR) model may not be suitable to fit this type of time series, especially when the time gaps are large or randomly distributed. We have proposed a set of models that extend the AR models for irregularly observed time series. Among the proposed models are the irregular Autoregressive (iAR) process, the Complex irregular Autoregressive (CiAR) process and the Bivariate irregular Autoregressive (BiAR) process. In addition, we have proposed extensions of the iAR process for Gamma and T distributed data. The functions implemented in this package allows for parameter estimation via maximum likelihood along with obtaining fitted values and forecasts. In this work we explain the commands of the iAR package and provide several illustrative examples of their use on both simulated and real-life data.