The midasr R package is dedicated to estimate restricted MIDAS regression, test whether the restriction conforms to data and use the estimated model for forecasting. The package is developed by Virmantas Kvedaras and Vaidotas Zemlys. The package development was influenced by features of the MIDAS Matlab toolbox created by Eric Ghysels. MIDAS-AR* specification code was written by Julius Vainora.
The package is available on CRAN. It can be installed in the usual way:
This will install the latest stable version, which currently is 0.6
The detailed description of the package features can be found in the JSS article. You can fine additional examples in the following file. The JSS article as .Rnw file and the accompanying files can be found in the midasr-user-guide github repository. Note that the version in github is slightly different from the one published.
The package has three demos:
logRV. Please run them with the following code:
library(midasr) demo("okun","midasr",ask=FALSE) demo("RV","midasr",ask=FALSE) demo("logRV","midasr",ask=FALSE)
The last 2 demos take some time to complete (about 5 minutes), so be patient.
To install the latest development version of the package use the following code:
The package versioning follows this convention: the stable version is x.y. This is a version which is available on CRAN. The development version is x.y.z and it can be installed from github. To install the specific version of the package use the following code:
If you find bugs or have suggestions please file an issue. Add the offending R code and the error message for a bug. Or write an email.