funtimes-package: funtimes: Functions for Time Series Analysis

Description Author(s) References


Advances in multiple aspects of time-series analysis are documented in this package. See available vignettes using
browseVignettes(package = "funtimes")

Tests for trends applicable to autocorrelated data, see
vignette("trendtests", package = "funtimes")
include bootstrapped versions of t-test and Mann–Kendall test \insertCiteNoguchi_etal_2011funtimes and bootstrapped version of WAVK test for possibly non-monotonic trends \insertCiteLyubchich_etal_2013_wavkfuntimes. The WAVK test is further applied in testing synchronism of trends \insertCiteLyubchich_Gel_2016_synchronismfuntimes; see an implementation to climate data in \insertCiteLyubchich_2016_trends;textualfuntimes. With iterative testing, the synchronism test is also applied for identifying clusters of multiple time series \insertCiteGhahari_etal_2017_MBDCEfuntimes.

Additional clustering methods are implemented using functions BICC \insertCiteSchaeffer_etal_2016_trustfuntimes and DR \insertCiteHuang_etal_2018_ridingfuntimes; function purity can be used to assess accuracy of clustering if true classes are known.

Changepoint detection methods include modified CUSUM-based bootstrapped test \insertCiteLyubchich_etal_2020_changepointsfuntimes.

Additional functions include implementation of the Beale's ratio estimator, see
vignette("beales", package = "funtimes")
Nonparametric comparison of tails of distributions is implemented using small bins defined based on quantiles \insertCiteSoliman_etal_2015_insurancefuntimes or intervals in the units in which the data are recorded \insertCiteLyubchich_Gel_2017_insurancefuntimes.

For a list of deprecated functions, use ?'funtimes-deprecated'


Maintainer: Vyacheslav Lyubchich (ORCID)


Other contributors:


funtimes documentation built on Nov. 28, 2020, 1:06 a.m.