Skip to content
exuber

Package

exuber: Econometric Analysis of Explosive Time Series

Also: exuber-package

exuber

Testing for and dating periods of explosive dynamics (exuberance) in time series using the univariate and panel recursive unit root tests proposed by Phillips et al. (2015) 10.1111/iere.12132 and Pavlidis et al. (2016) 10.1007/s11146-015-9531-2. The recursive least-squares algorithm uses the matrix inversion lemma, so no matrix has to be inverted at each step, which makes the tests much faster to compute. The package also simulates a variety of periodically collapsing bubble processes. Details can be found in Vasilopoulos et al. (2022) 10.18637/jss.v103.i10.

Package options

exuber.show_progress

  • Should lengthy operations such as radf_mc_cv() show a progress bar? Default: TRUE

exuber.parallel

  • Should lengthy operations use parallel computation? Default: TRUE in an interactive session and FALSE otherwise (scripts, knitr and R CMD check), because starting workers costs a few seconds. Set it to TRUE in a script to opt in. The worker cluster is started once per session and reused. The radf_*_cv() and radf_*_distr() simulation engines honor the option (radf_mc_cv(), radf_wb_cv(), radf_sb_cv(), radf_recovery_cv() and radf_common_cv()). The standalone tests and monitors run serially regardless.

exuber.ncores

  • How many cores to use for parallel computation. Default: the number of system cores minus 1 (2 in a non-interactive session), capped by the MC_CORES environment variable when it is set.

exuber.global_seed

  • When set, the seed feeds automatically into all functions that generate random numbers. Default: NA

See also

Useful links: