Package: polle 1.6.4

polle: Policy Learning

Package for learning and evaluating (subgroup) policies via doubly robust loss functions. Policy learning methods include doubly robust blip/conditional average treatment effect learning and sequential policy tree learning. Methods for (subgroup) policy evaluation include doubly robust cross-fitting and online estimation/sequential validation. See Nordland and Holst (2026) <doi:10.18637/jss.v116.i04> for documentation and references.

Authors:Andreas Nordland [aut, cre], Klaus Holst [aut]

polle_1.6.4.tar.gz
polle_1.6.4.zip(r-4.7)polle_1.6.4.zip(r-4.6)polle_1.6.4.zip(r-4.5)
polle_1.6.4.tgz(r-4.6-any)polle_1.6.4.tgz(r-4.5-any)
polle_1.6.4.tar.gz(r-4.7-any)polle_1.6.4.tar.gz(r-4.6-any)
polle_1.6.4.tgz(r-4.6-emscripten)
manual.pdf |manual.html✨
DESCRIPTION |NEWS
card.svg |card.png
polle/json (API)

# Install 'polle' in R:
install.packages('polle', repos = c('https://andreasnordland.r-universe.dev', 'https://cloud.r-project.org'))

Bug tracker:https://github.com/andreasnordland/polle/issues

On CRAN:

Conda:

7.01 score 6 stars 13 scripts 2.5k downloads 55 exports 51 dependencies

Last updated from:f92bfd5f61. Checks:7 ERROR, 2 OK. Indexed: yes.

TargetResultTimeFilesSyslog
linux-devel-x86_64ERROR148
source / vignettesOK259
linux-release-x86_64ERROR141
macos-release-arm64ERROR128
macos-oldrel-arm64ERROR140
windows-develERROR82
windows-releaseERROR82
windows-oldrelERROR312
wasm-releaseOK177

Exports:Allc_coxc_no_censoringconditionalcontrol_blipcontrol_drqlcontrol_earlcontrol_owlcontrol_ptlcontrol_rwlcopy_policy_dataestimatefit_c_functionsfit_g_functionsg_empirg_glmg_glmnetg_rfg_slg_xgboostget_action_setget_actionsget_eventget_g_functionsget_historyget_history_namesget_idget_id_stageget_Kget_nget_policyget_policy_actionsget_policy_functionsget_policy_objectget_q_functionsget_stage_action_setsget_utilityICpartialpolicy_datapolicy_defpolicy_evalpolicy_eval_onlinepolicy_learnq_glmq_glmnetq_rfq_slq_xgboostsim_multi_stagesim_single_stagesim_single_stage_multi_actionssim_two_stagesim_two_stage_multi_actionssubset_id

Dependencies:abindBHbitopscaToolsclicodetoolscvAUCdata.tabledfoptimDiceKrigingdigestDynTxRegimeforeachfuturefuture.applygamglobalsgplotsgrfgtoolsiteratorskernlabKernSmoothlatticelavalistenvlmtestMatrixmetsmodelObjmvtnormnnlsnumDerivparallellypolicytreeprogressrquadprogR6RcppRcppArmadilloRcppEigenrgenoudrlangROCRsandwichSQUAREMSuperLearnersurvivaltargetedtimeregzoo

Right censoring/monotone coarsening
Target parameter under discrete time right-censoring/monotone coarsening | Efficient estimation | Adaption to continuous time right-censoring | Terminal events | Examples | References
Rendered fromright_censoring.Rmdusingknitr::rmarkdown

Last update: 2025-11-05
Started: 2025-10-30

policy_data
Single-stage: wide data | Two-stage: wide data | Multi-stage: long data | SessionInfo | References
Rendered frompolicy_data.Rmdusingknitr::rmarkdown

Last update: 2025-10-30
Started: 2024-04-23

Estimating and Evaluating the Optimal Subgroup
Setup | Threshold policy learning | Subgroup average treatment effect | Asymptotics | References
Rendered fromoptimal_subgroup.Rmdusingknitr::rmarkdown

Last update: 2024-07-29
Started: 2024-07-11

policy_eval
Evaluating a user-defined policy | Working with policy_eval objects | Nuisance models | Evaluating a policy learning algorithm | Cross-fitting | Parallel processing via future.apply | SessionInfo | References
Rendered frompolicy_eval.Rmdusingknitr::rmarkdown

Last update: 2024-07-26
Started: 2024-04-23

policy_learn
Specifying and applying a policy learner | Cross-fitting the doubly robust score | Realistic policy learning | Implementation/Simulation and get_policy_functions() | Policy objects and get_policy() | SessionInfo | References
Rendered frompolicy_learn.Rmdusingknitr::rmarkdown

Last update: 2024-04-23
Started: 2024-04-23

Readme and manuals

Help Manual

Help pageTopics
c_model class objectc_cox c_model c_no_censoring
Conditional Policy Evaluationconditional
Control arguments for doubly robust blip-learningcontrol_blip
Control arguments for doubly robust Q-learningcontrol_drql
Control arguments for Efficient Augmentation and Relaxation Learningcontrol_earl
Control arguments for Outcome Weighted Learningcontrol_owl
Control arguments for Policy Tree Learningcontrol_ptl
Control arguments for Residual Weighted Learningcontrol_rwl
Copy Policy Data Objectcopy_policy_data
Fit Censoring Functionsfit_c_functions
Fit g-functionsfit_g_functions
g_model class objectg_empir g_glm g_glmnet g_model g_rf g_sl g_xgboost
Get Action Setget_action_set
Get Actionsget_actions
Get event indicatorsget_event
Get g-functionsget_g_functions
Get history variable namesget_history_names
Get IDsget_id
Get IDs and Stagesget_id_stage
Get Maximal Stagesget_K
Get Number of Observationsget_n
Get Policyget_policy
Get Policy Actionsget_policy_actions
Get Policy Functionsget_policy_functions get_policy_functions.blip get_policy_functions.drql get_policy_functions.ptl get_policy_functions.ql
Get Policy Objectget_policy_object
Get Q-functionsget_q_functions
Get Stage Action Setsget_stage_action_sets
Get the Utilityget_utility
Get History Objectget_history history
Nuisance Functionsnuisance_functions
Trim Number of Stagespartial
Plot policy data for given policiesplot.policy_data
Plot histogram of the influence curve for a 'policy_eval' objectplot.policy_eval
Policy-classpolicy
Create Policy Data Objectpolicy_data print.policy_data summary.policy_data
Define Policypolicy_def
Policy Evaluation+.policy_eval coef.policy_eval estimate.policy_eval IC.policy_eval merge.policy_eval policy_eval print.policy_eval vcov.policy_eval vcov.policy_eval_online
Online/Sequential Policy Evaluationpolicy_eval_online
Create Policy Learnerpolicy_learn policy_object print.policy_learn print.policy_object
Predict g-functions and Q-functionspredict.nuisance_functions
q_model class objectq_glm q_glmnet q_model q_rf q_sl q_xgboost
Simulate Multi-Stage Datasim_multi_stage
Simulate Single-Stage Datasim_single_stage
Simulate Single-Stage Multi-Action Datasim_single_stage_multi_actions
Simulate Two-Stage Datasim_two_stage
Simulate Two-Stage Multi-Action Datasim_two_stage_multi_actions
Subset Policy Data on IDsubset_id