Simulation and Analysis of Contextual Multi-Armed Bandit Policies

[Up] [Top]

Documentation for package ‘contextual’ version

Help Pages


-- A --

Agent Agent
agent Agent
arms Plot
average Plot

-- B --

Bandit Bandit: Superclass
bandit Bandit: Superclass
BasicBernoulliBandit Bandit: BasicBernoulliBandit
BasicGaussianBandit Bandit: BasicGaussianBandit
BootstrapTSPolicy Policy: Thompson sampling with the online bootstrap

-- C --

check_history_data Plot
clear_data_table History
clipr Clip vectors
ContextualBernoulliBandit Bandit: Naive Contextual Bernouilli Bandit
ContextualBinaryBandit Bandit: ContextualBinaryBandit
ContextualEpochGreedyPolicy Policy: A Time and Space Efficient Algorithm for Contextual Linear Bandits
ContextualEpsilonGreedyPolicy Policy: ContextualEpsilonGreedyPolicy with unique linear models
ContextualHybridBandit Bandit: ContextualHybridBandit
ContextualLinearBandit Bandit: ContextualLinearBandit
ContextualLinTSPolicy Policy: Linear Thompson Sampling with unique linear models
ContextualLogitBandit Bandit: ContextualLogitBandit
ContextualLogitBTSPolicy Policy: ContextualLogitBTSPolicy
ContextualPrecachingBandit Bandit: ContextualPrecachingBandit
ContextualTSProbitPolicy Policy: ContextualTSProbitPolicy
ContextualWheelBandit Bandit: ContextualWheelBandit
ContinuumBandit Bandit: ContinuumBandit
cumulative History

-- D --

data_table_factors_to_numeric Convert all factor columns in data.table to numeric
dec<- Decrement
dinvgamma The Inverse Gamma Distribution
do_plot Plot
do_step Agent

-- E --

EpsilonFirstPolicy Policy: Epsilon First
EpsilonGreedyPolicy Policy: Epsilon Greedy
Exp3Policy Policy: Exp3

-- F --

FixedPolicy Policy: Fixed Arm
formatted_difftime Format difftime objects

-- G --

generate_bandit_data Bandit: Superclass
get_action Policy: Superclass
get_arm_context Return context vector of an arm
get_context Bandit: Superclass
get_data_frame History
get_data_table History
get_full_context Get full context matrix over all arms
get_global_seed Lookup .Random.seed in global environment
get_t Agent
gg_color_hue Plot
gittinsbrezzilai Policy: Gittins Approximation algorithm for choosing arms in a MAB problem.
GittinsBrezziLaiPolicy Policy: Gittins Approximation algorithm for choosing arms in a MAB problem.
GradientPolicy Policy: Gradient

-- H --

History History

-- I --

inc<- Increment
ind On-the-fly indicator function for use in formulae
initialize_theta Policy: Superclass
inv Inverse from Choleski (or QR) Decomposition.
invgamma The Inverse Gamma Distribution
invlogit Inverse Logit Function
is_rstudio Check if in RStudio

-- L --

LifPolicy Policy: Continuum Bandit Policy with Lock-in Feedback
LinUCBDisjointOptimizedPolicy Policy: LinUCB with unique linear models
LinUCBDisjointPolicy Policy: LinUCB with unique linear models
LinUCBGeneralPolicy Policy: LinUCB with unique linear models
LinUCBHybridOptimizedPolicy Policy: LinUCB with hybrid linear models
LinUCBHybridPolicy Policy: LinUCB with hybrid linear models
load History

-- M --

mvrnorm Simulate from a Multivariate Normal Distribution

-- O --

OfflineBootstrappedReplayBandit Bandit: Offline Bootstrapped Replay
OfflineDirectMethodBandit Bandit: Offline Direct Methods
OfflineDoublyRobustBandit Bandit: Offline Doubly Robust
OfflineLookupReplayEvaluatorBandit Bandit: Offline Replay with lookup tables
OfflinePropensityWeightingBandit Bandit: Offline Propensity Weighted Replay
OfflineReplayEvaluatorBandit Bandit: Offline Replay
ones_in_zeroes A vector of zeroes and ones
one_hot One Hot Encoding of data.table columns
optimal Plot
OraclePolicy Policy: Oracle

-- P --

pinvgamma The Inverse Gamma Distribution
Plot Plot
plot.History Plot Method for Contextual History
plot.history Plot Method for Contextual History
Policy Policy: Superclass
policy Policy: Superclass
post_initialization Bandit: Superclass
print.History Print Method for Contextual History
print.history Print Method for Contextual History
print_data History
prob_winner Binomial Win Probability

-- Q --

qinvgamma The Inverse Gamma Distribution

-- R --

RandomPolicy Policy: Random
rinvgamma The Inverse Gamma Distribution
run Simulator

-- S --

sample_one_of Sample one element from vector or list
save History
set_data_frame History
set_data_table History
set_external Change Default Graphing Device from RStudio
set_global_seed Set .Random.seed to a pre-saved value
set_parameters Policy: Superclass
set_reward Policy: Superclass
set_t Agent
sherman_morrisson Sherman-Morrisson inverse
Simulator Simulator
simulator Simulator
sim_post Binomial Posterior Simulator
SoftmaxPolicy Policy: Softmax
summary.History Summary Method for Contextual History
summary.history Summary Method for Contextual History
sum_of Sum of list

-- T --

theta Policy: Superclass
ThompsonSamplingPolicy Policy: Thompson Sampling

-- U --

UCB1Policy Policy: UCB1
UCB2Policy Policy: UCB2

-- V --

value_remaining Potential Value Remaining
var_welford Welford's variance

-- W --

which_max_list Get maximum value in list
which_max_tied Get maximum value randomly breaking ties