whatifbandit0.3.0 package

Calculate Adaptive AIPW Estimates

Adaptively Assign Treatments in a Period

Checks Validity of Thompson sampling probabilities

Checking For Valid Assignment Methods

Checking existence and declaration of columns

Checking for Valid Input Data

Checking Imputation Info

Check Level

Checking if Inputs are Logical Values (TRUE and FALSE)

Checking If Inputs Are Positive Integers or a Valid String

Checking for Proportions

Column arguments shared across functions

Condenses results into a list for multiple_mab_simulation()

Creating proper conditions vector

Create Treatment Wave Cutoffs

Create Necessary Columns for Multi-Arm Bandit Trial

Create Prior Periods

Ends Multi-Arm Bandit Trial

Calculates Number of Observations Assigned to Each Treatment

Calculate Multi-Arm Bandit Decision Based on Algorithm

Thompson sampling Algorithm

UCB1 Sampling Algorithm

Calculate Observation Level AIPW For Each Treatment Condition

Gather Past Results for Given Assignment Period

Precomputing Key Values for Outcome Imputation

Outcome Imputation Preparation

Imputing New Outcomes of Multi-Arm-Bandit Trial

Simulates Multi-Arm Bandit Trial From Prepared Inputs

Run Multiple Multi-Arm-Bandit Trials with Inference in Parallel

Plot Treatment Arms Over Time

Plot Cumulative Assignment Probability Over Time

Plot AIPW Estimates

Plots Histograms of multiple_mab_simulation() Results

Plots AIPW Confidence Intervals

Plot Treatment Arms Over Multiple Trials

Plot Generic for mab objects

Plot Generic For multiple.mab Objects

Pre-Simulation Setup for an adaptive Multi-Arm-Bandit Trial

Print Helper for mab and multiple.mab

Print Generic For mab

Print Generic For multiple.mab

Runs Multi-Arm Bandit Trial

Run One Adaptive Simulation With Inference.

Summary Generic For "mab" Class

Summary Generic For "multiple.mab" Class

Validates Inputs For single_mab_simulation() and `multiple_mab_simul...

Verbose Printer

whatifbandit: Analyzing Randomized Experiments as Multi-Arm Bandits