Many of the examples in the first three chapters assume independent arms: G = F 1 × ... × F k. In this chapter we consider independent Bernoulli arms. As has been our convention in the Bernoulli case, we regard F i as a distribution on the Bernoulli parameter θ i ∈ [0, 1] rather than on Q i ∈ D; and consistent with an earlier modification of notation, we write the conditional distribution of (θ 1, θ 2,...,θ k ) given success on arm 1, say, as $$ {\sigma _1}\left( {{F_1},{F_2},...,{F_k}} \right) = \left( {\sigma {F_1},{F_2},...,{F_k}} \right), $$ and given a failure on arm 1 and as $$ {\varphi _1}\left( {{F_1},{F_2},...,{F_k}} \right) = \left( {\varphi {F_1},{F_2},...,{F_k}} \right). $$
CITATION STYLE
Berry, D. A., & Fristedt, B. (1985). Independent Bernoulli arms. In Bandit problems (pp. 65–82). Springer Netherlands. https://doi.org/10.1007/978-94-015-3711-7_4
Mendeley helps you to discover research relevant for your work.