Shared-control multi-arm designs | Stratified designs | Raw unbounded outcomes
Package: cmrdesign 0.1.0
cmrdesign: Conditional Minimax Regret Design Rules
Implements Conditional Minimax Regret design rules for pilot-informed experimental assignment, as proposed in Yamin (2026) <doi:10.48550/arXiv.2607.16982>. Includes two-arm, unbounded two-arm, multi-arm, stratified, multiple-outcome, proxy-outcome, and pilot-planning tools. Variance bounds include Maurer and Pontil (2009) <doi:10.48550/arXiv.0907.3740> empirical-Bernstein bounds and exact Bernoulli confidence bounds.
Authors:
cmrdesign_0.1.0.tar.gz
cmrdesign_0.1.0.zip(r-4.7-any)cmrdesign_0.1.0.zip(r-4.6-any)cmrdesign_0.1.0.zip(r-4.5-any)
cmrdesign_0.1.0.tgz(r-4.6-any)cmrdesign_0.1.0.tgz(r-4.5-any)
cmrdesign_0.1.0.tar.gz(r-4.7-any)cmrdesign_0.1.0.tar.gz(r-4.6-any)
cmrdesign_0.1.0.tgz(r-4.6-emscripten)
manual.pdf |manual.html✨
DESCRIPTION |NEWS
card.svg |card.png
cmrdesign/json (API)
| # Install 'cmrdesign' in R: |
| install.packages('cmrdesign', repos = c('https://juancyamin.r-universe.dev', 'https://cloud.r-project.org')) |
Bug tracker:https://github.com/juancyamin/cmrdesign/issues
Pkgdown/docs site:https://juancyamin.github.io
causal-inferencecmrconditional-minimax-regretexperimental-designpilot-studiespython-packagerandomized-experimentssample-size-planningtreatment-assignment
Last updated from:bdcec68f5f. Checks:9 OK. Indexed: yes.
| Target | Result | Time | Files | Syslog |
|---|---|---|---|---|
| linux-devel | OK | 128 | ||
| source / vignettes | OK | 180 | ||
| linux-release | OK | 122 | ||
| macos-release | OK | 190 | ||
| macos-oldrel | OK | 148 | ||
| windows-devel | OK | 86 | ||
| windows-release | OK | 82 | ||
| windows-oldrel | OK | 84 | ||
| wasm-release | OK | 91 |
Exports:activation_threshold_bernoulliactivation_threshold_boundedassign_additive_regularized_neymanassign_balanceassign_exponential_regularized_neymanassign_feasible_neymanassign_multiarm_balanceassign_multiarm_neymanassign_neymanassign_stratified_balanceassign_stratified_neymanassign_trimmed_neymanbinary_rectangle_cornersbinary_rectangle_regretboundary_indicatorboundary_ratebreak_even_pilot_sharecertificate_validcmr_binarycmr_binary_from_rectanglecmr_delayed_outcomecmr_multiarmcmr_multiarm_from_rectanglecmr_multiple_outcomescmr_plancmr_proxycmr_stratifiedcmr_stratified_from_rectanglecmr_two_armcmr_two_arm_from_rectanglecmr_unboundedcmr_unbounded_from_rectanglecoverage_indicatorfolded_binomial_pmffolded_binomial_tailsmultiarm_oracle_variancemultiarm_rectangle_verticesmultiarm_regretmultiarm_variance_objectiveoracle_variancepilot_planpilot_viability_bandrealize_allocationrectangle_bernoulli_binaryrectangle_bernoulli_two_armrectangle_binaryrectangle_bounded_binaryrectangle_bounded_two_armrectangle_delayed_outcomerectangle_multiarmrectangle_multiple_outcomesrectangle_proxyrectangle_stratifiedrectangle_two_armrectangle_unboundedregretsaving_vs_balanceshare_of_oracle_gainstratified_oracle_variancestratified_rectangle_verticesstratified_regretstratified_variance_objectivevariance_bounds_bernoulli_exactvariance_bounds_martinez_taboada_ramdasvariance_bounds_maurer_pontilvariance_bounds_unbounded_momvariance_objective
Dependencies:
CMR Extensions
Last update: 2026-07-25
Started: 2026-07-23
Two-Arm Bounded Outcomes
Simulate a pilot | Compute the CMR assignment | Convert the share to integer counts | Compare confidence-set methods | Work from a precomputed rectangle
Last update: 2026-07-25
Started: 2026-07-23
Multiple Outcomes
Simulate a pilot with two outcomes | Weighted-index CMR | Co-primary Outcome CMR | Compare index and co-primary designs
Last update: 2026-07-24
Started: 2026-07-23
Binary Outcomes
Simulate a binary pilot | Exact Bernoulli CMR | Inspect the folded-binomial sufficient statistic
Last update: 2026-07-23
Started: 2026-07-23
Pilot Planning
Planning inputs | Break-even pilot share | Activation thresholds | Viability band | Check a desired pilot size
Last update: 2026-07-23
Started: 2026-07-23
Proxy Outcomes
Simulate proxy and primary outcomes | Choose a bridge constant | Proxy CMR | Simulated oracle comparison
Last update: 2026-07-23
Started: 2026-07-23
Readme and manuals
Help Manual
| Help page | Topics |
|---|---|
| cmrdesign: Conditional Minimax Regret design rules | cmrdesign-package cmrdesign |
| Pilot-size activation thresholds | activation_threshold_bernoulli activation_threshold_bounded |
| Baseline and comparator assignment rules | assign_additive_regularized_neyman assign_balance assign_exponential_regularized_neyman assign_feasible_neyman assign_multiarm_balance assign_stratified_balance assign_trimmed_neyman |
| Two-arm CMR from a variance rectangle | binary_rectangle_corners binary_rectangle_regret cmr_binary_from_rectangle cmr_two_arm_from_rectangle |
| Break-even pilot share | break_even_pilot_share |
| Shared-control multi-arm CMR assignment | cmr_multiarm |
| Multi-arm CMR from a variance rectangle | cmr_multiarm_from_rectangle |
| Multiple-outcome CMR assignment | cmr_multiple_outcomes |
| Proxy or delayed-outcome CMR assignment | cmr_delayed_outcome cmr_proxy |
| Stratified CMR assignment | cmr_stratified |
| Stratified CMR from a variance rectangle | cmr_stratified_from_rectangle |
| Two-arm Conditional Minimax Regret assignment | cmr_binary cmr_two_arm |
| Unbounded-outcome CMR assignment | cmr_unbounded |
| Unbounded-outcome CMR from a variance rectangle | cmr_unbounded_from_rectangle |
| Diagnostic metrics for CMR simulations | boundary_indicator boundary_rate certificate_valid coverage_indicator saving_vs_balance share_of_oracle_gain |
| Folded-binomial distribution for Bernoulli sample variances | folded_binomial_pmf folded_binomial_tails |
| Multi-arm variance objectives and Neyman allocation | assign_multiarm_neyman multiarm_oracle_variance multiarm_rectangle_vertices multiarm_regret multiarm_variance_objective |
| Pilot/main-wave planning summary | cmr_plan pilot_plan |
| Pilot viability band | pilot_viability_band |
| Print and summarize CMR results | print.cmr_multiarm print.cmr_multiple_outcomes print.cmr_proxy print.cmr_stratified print.cmr_two_arm print.cmr_unbounded print.summary.cmr_result summary.cmr_multiarm summary.cmr_multiple_outcomes summary.cmr_proxy summary.cmr_stratified summary.cmr_two_arm summary.cmr_unbounded |
| Convert CMR target shares to integer allocation counts | print.cmr_allocation realize_allocation |
| Two-arm confidence rectangles | rectangle_bernoulli_binary rectangle_bernoulli_two_arm rectangle_binary rectangle_two_arm |
| Bounded two-arm confidence rectangle | rectangle_bounded_binary rectangle_bounded_two_arm |
| Multi-arm confidence rectangle | rectangle_multiarm |
| Multiple-outcome confidence rectangle | rectangle_multiple_outcomes |
| Proxy or delayed-outcome confidence rectangle | rectangle_delayed_outcome rectangle_proxy |
| Stratified confidence rectangle | rectangle_stratified |
| Unbounded-outcome confidence rectangle | rectangle_unbounded |
| Stratified variance objectives and Neyman allocation | assign_stratified_neyman stratified_oracle_variance stratified_rectangle_vertices stratified_regret stratified_variance_objective |
| Exact Bernoulli variance bounds | variance_bounds_bernoulli_exact |
| Bounded-outcome variance bounds | variance_bounds_martinez_taboada_ramdas variance_bounds_maurer_pontil |
| Unbounded-outcome median-of-means variance bounds | variance_bounds_unbounded_mom |
| Two-arm variance objectives and Neyman allocation | assign_neyman oracle_variance regret variance_objective |