Functions for Rare Events Analysis R package rarefun version 0.1.0 Functions for detecting and analyzing rare events in data. Implements isolation forest (Liu et al., 2008, <doi:10.1109/ICDM.2008.17>) and clustering for anomaly detection in time series residuals. Decomposes time series using LOESS (Locally Estimated Scatterplot Smoothing) or STL (Seasonal-Trend decomposition using LOESS). Detects marine heatwaves and cold spells following Hobday et al. (2016) <doi:10.1016/j.pocean.2015.12.014>. Provides goodness-of-fit tests for quantile regression (Haupt et al., 2011, <doi:10.1080/02664763.2011.573542>), partial dependence with quantile random forests, MCC (Matthews Correlation Coefficient) computation and testing, knee-point detection via the Kneedle algorithm (Satopaa et al., 2011, <doi:10.1109/ICDCSW.2011.20>), and spatial point matching. Software Depends: R (>= 4.1.0) Imports: RANN, boot, dbscan, geosphere, isotree, Rdpack, parallel, pdp, stats Suggests: testthat (>= 3.0.0), dplyr, ggplot2, heatwaveR, patchwork, quantmod, quantreg, ranger, viridis Vyacheslav Lyubchich <lyubchich@umces.edu> Comprehensive R Archive Network (CRAN) Vyacheslav Lyubchich [aut, cre] (ORCID: <https://orcid.org/0000-0001-7936-4285>), Geneviève Nesslage [aut] (ORCID: <https://orcid.org/0000-0003-1770-6803>) MIT + file LICENSE (https://CRAN.R-project.org/package=rarefun/LICENSE) 2026-06-02 application/tgz https://CRAN.R-project.org/package=rarefun doi:10.32614/CRAN.package.rarefun