The code for the following sections

- "2. Getting started",
- "2.1. Additional features",
- "3.1. E-greedy", and
- "3.2. Thompson sampling for the K-armed Bernoulli bandit"
- "3.3. Thompson sampling for optimal design"

is supplied with the package StreamingBandit in the default experiments.

***

o The arguments to be defined to reproduce the results of
  "2. Getting started" are
  - N = 150, 
  - seed = 1271246, and
  - verbose = FALSE
  ... as reported on page 13 of the manuscript.

o The arguments to be defined to reproduce the results of
  "3.1. E-greedy" and 
  "3.2. Thompson sampling for the K-armed Bernoulli bandit"
  are
  - N = 1000, 
  - seed = 1271246, and
  - verbose = FALSE
  ... as reported on page 18/19 of the manuscript.

o The arguments to be defined to reproduce the results of
  "3.3. Thompson sampling for optimal design" are
  - N = 1000, 
  - seed = 43123, and
  - verbose = FALSE
  ... as reported on page 21 of the manuscript.



