Package: seqcomp 0.1.0

Akbar Alasgarli

seqcomp: Sequential Comparison of Probabilistic Forecasts

Implements tools for sequential comparison of probabilistic forecasts, including binary and categorical scoring rules, anytime-valid confidence sequences, e-processes, Winkler-score comparisons, lag handling, and predictable-bound e-processes.

Authors:Akbar Alasgarli [aut, cre]

seqcomp_0.1.0.tar.gz
seqcomp_0.1.0.tar.gz(r-4.7-any)seqcomp_0.1.0.tar.gz(r-4.6-any)
seqcomp_0.1.0.tgz(r-4.6-emscripten)
manual.pdf |manual.html✨
DESCRIPTION |NEWS
card.svg |card.png
seqcomp/json (API)

# Install 'seqcomp' in R:

install.packages('seqcomp', repos = c('https://cran.r-universe.dev', 'https://cloud.r-project.org'))

Bug tracker:https://github.com/alasgarliakbar/seqcomp/issues

On CRAN:

2.70 score 4 scripts 30 exports 3 dependencies

Last updated from:55934cbcba. Checks:4 OK. Indexed: yes.

Target	Result	Time
linux-devel-x86_64	OK	128
source / vignettes	OK	173
linux-release-x86_64	OK	126
wasm-release	OK	163

Exports:brier_score calibrate_p_to_e cm_boundary compare_forecasts crps_empirical crps_normal crps_std cs_asymptotic cs_bernstein cs_hoeffding eprocess eprocess_lag eprocess_predictable eprocess_rejections ge_boundary log_score predictable_rejections ps_boundary qlike_score rho_from_vopt score_bounds score_diff_scales spherical_score split_streams tick_loss unroll_stream winkler_compare winkler_cs winkler_etest winkler_score

Dependencies:lamW Rcpp RcppParallel

Getting started with seqcomp

Rendered fromseqcomp.Rmdusingknitr::rmarkdown

Last update: 2026-06-30
Started: 2026-06-30

Citation

Readme and manuals

Help Manual

Help page	Topics
seqcomp: Sequential Comparison of Probabilistic Forecasts	seqcomp-package seqcomp
Brier score for binary and categorical forecasts	brier_score
P-to-e calibrator	calibrate_p_to_e
Normal mixture (CM) boundary	cm_boundary
Compare Two Sequential Forecasters	compare_forecasts
Negated CRPS for empirical predictive distributions	crps_empirical
Negated CRPS for normal predictive distributions	crps_normal
Negated CRPS for Student-t predictive distributions	crps_std
Asymptotic confidence sequence (Appendix C, Eq. 55, Choe & Ramdas 2023)	cs_asymptotic
Empirical Bernstein confidence sequence (Theorem 2, Choe & Ramdas 2023)	cs_bernstein
Hoeffding-style confidence sequence (Theorem 1, Choe & Ramdas 2023)	cs_hoeffding
Sub-exponential mixture e-process (Theorem 3, Choe & Ramdas 2023)	eprocess
Lag-h e-process for sequential forecast comparison (Propositions 5 & 6)	eprocess_lag
Fixed-lambda e-process with predictable bounds (Proposition 7)	eprocess_predictable
Determine rejection times for an e-process output	eprocess_rejections
Gamma-exponential mixture boundary	ge_boundary
Logarithmic score for binary and categorical forecasts	log_score
Summarise predictable bounds e-process	predictable_rejections
Polynomial stitched (PS) boundary	ps_boundary
Negated QLIKE score for variance forecasts	qlike_score
Convert optimal intrinsic time to rho	rho_from_vopt
Score difference bounds for a named scoring rule	score_bounds
Score difference bounds -> sub-Gaussian / sub-exponential scale	score_diff_scales
Spherical score for binary and categorical forecasts	spherical_score
Split a sequence into h interleaved lag streams	split_streams
Negated tick loss for quantile forecasts	tick_loss
Unroll a stream-wise quantity back to the original time scale	unroll_stream
Full Winkler comparison pipeline (Proposition 4)	winkler_compare
One-sided empirical Bernstein CS for Winkler scores (Proposition 4)	winkler_cs
E-process for Winkler scores (Proposition 4 + Theorem 3)	winkler_etest
Winkler-normalized binary score	winkler_score

Package: seqcomp 0.1.0

seqcomp: Sequential Comparison of Probabilistic Forecasts

Citation

Readme and manuals

Help Manual

Usage by other packages (reverse dependencies)