Package: banditsCI 1.0.0

Molly Offer-Westort

banditsCI: Bandit-Based Experiments and Policy Evaluation

Frequentist inference on adaptively generated data. The methods implemented are based on Zhan et al. (2021) <doi:10.48550/arXiv.2106.02029> and Hadad et al. (2021) <doi:10.48550/arXiv.1911.02768>. For illustration, several functions for simulating non-contextual and contextual adaptive experiments using Thompson sampling are also supplied.

Authors:Molly Offer-Westort [aut, cre, cph], Yinghui Zhou [aut], Ruohan Zhan [aut]

banditsCI_1.0.0.tar.gz
banditsCI_1.0.0.zip(r-4.7-any)banditsCI_1.0.0.zip(r-4.6-any)banditsCI_1.0.0.zip(r-4.5-any)
banditsCI_1.0.0.tgz(r-4.6-any)banditsCI_1.0.0.tgz(r-4.5-any)
banditsCI_1.0.0.tar.gz(r-4.7-any)banditsCI_1.0.0.tar.gz(r-4.6-any)
banditsCI_1.0.0.tgz(r-4.6-emscripten)
manual.pdf |manual.html✨
DESCRIPTION |NEWS
card.svg |card.png
banditsCI/json (API)

# Install 'banditsCI' in R:

install.packages('banditsCI', repos = c('https://uchicago-pol-methods.r-universe.dev', 'https://cloud.r-project.org'))

Bug tracker:https://github.com/uchicago-pol-methods/banditsci/issues

Pkgdown/docs site:https://uchicago-pol-methods.github.io

On CRAN:

4.48 score 3 stars 10 scripts 313 downloads 24 exports 14 dependencies

Last updated from:ba5e4ff318. Checks:9 OK. Indexed: yes.

Target	Result	Time
linux-devel	OK	105
source / vignettes	OK	297
linux-release	OK	110
macos-release	OK	173
macos-oldrel	OK	177
windows-devel	OK	84
windows-release	OK	86
windows-oldrel	OK	78
wasm-release	OK	89

Exports:.check_A .check_first_batch .check_shape aw_estimate aw_scores aw_var calculate_balwts calculate_continuous_X_statistics draw_thompson estimate generate_bandit_data ifelse_clip impose_floor LinTSModel output_estimates plot_cumulative_assignment ridge_init ridge_muhat_lfo_pai ridge_update run_experiment simple_tree_data stick_breaking twopoint_stable_var_ratio update_thompson

Dependencies:codetools foreach glmnet iterators lattice MASS Matrix mvtnorm rbibutils Rcpp RcppEigen Rdpack shape survival

Confidence interval vignette.

Set parameters. | For the contextual case. | Estimating response. | For the non-contextual case.

Rendered frombanditsCI.Rmdusingknitr::rmarkdown

Last update: 2024-12-01
Started: 2023-09-08

Citation

Development and contributors

Readme and manuals

Help Manual

Help page	Topics
Estimate policy value via non-contextual adaptive weighting.	aw_estimate
Compute AIPW/doubly robust scores.	aw_scores
Variance of policy value estimator via non-contextual adaptive weighting.	aw_var
Calculate balancing weight scores.	calculate_balwts
Estimate/variance of policy evaluation via contextual weighting.	calculate_continuous_X_statistics
Thompson Sampling draws.	draw_thompson
Estimate/variance of policy evaluation via non-contextual weighting.	estimate
Generate classification data.	generate_bandit_data
Clip lamb values between a minimum x and maximum y.	ifelse_clip
Impose probability floor.	impose_floor
Linear Thompson Sampling model.	LinTSModel
Policy evaluation with adaptively generated data.	output_estimates
Plot cumulative assignment for bandit experiment.	plot_cumulative_assignment
Ridge Regression Initialization for Arm Expected Rewards	ridge_init
Leave-future-out ridge-based estimates for arm expected rewards.	ridge_muhat_lfo_pai
Updates ridge regression matrices.	ridge_update
Run an experiment using Thompson Sampling.	run_experiment
Generate simple tree data.	simple_tree_data
Stick breaking function.	stick_breaking
Calculate allocation ratio for a two-point stable-variance bandit.	twopoint_stable_var_ratio
Update linear Thompson Sampling model.	update_thompson