Implementation of robust and non-robust Standard Normal Homogeneity Test (SNHT) for changepoint detection. This test statistic is equal sided, as proposed in "Homogenization of Radiosonde Temperature Time Series Using Innovation Statistics" by Haimberger, L., 2007. However, the statistic contains an estimate of sigma^2 in the denominator instead of sigma, which seems to be a more appropriate value (based on the paper "Homogenization of Swedish temperature data. Part I: Homogeneity test for linear trends." by Alexandersson, H., and A. Moberg, 1997).
|Imports:||mgcv, ggplot2, gridExtra, zoo, plyr, reshape2, methods|
|Author:||Josh Browning [aut, cre], Carina Schneider [aut]|
|Maintainer:||Josh Browning <rockclimber112358 at gmail.com>|
|CRAN checks:||snht results|
snht: Robust and non-robust SNHT tests for changepoint detection.
|Windows binaries:||r-devel: snht_1.0.5.zip, r-release: snht_1.0.5.zip, r-oldrel: snht_1.0.5.zip|
|macOS binaries:||r-release (arm64): snht_1.0.5.tgz, r-release (x86_64): snht_1.0.5.tgz, r-oldrel: snht_1.0.5.tgz|
|Old sources:||snht archive|
Please use the canonical form https://CRAN.R-project.org/package=snht to link to this page.