value_remaining {bandit} | R Documentation |
Compute the "value_remaining" in the binomial bandits
value_remaining(x, n, alpha = 1, beta = 1, ndraws = 10000)
x |
as in prop.test, a vector of the number of successes |
n |
as in prop.test, a vector of the number of trials |
alpha |
shape parameter alpha for the prior beta distribution. |
beta |
shape parameter beta for the prior beta distribution. |
ndraws |
number of random draws from the posterior |
value_remaining distribution; the distribution of improvement amounts that another arm might have over the current best arm
Thomas Lotze and Markus Loecher
https://support.google.com/analytics/answer/2846882?hl=en&topic=2844866&rd=1
x=c(10,20,30,80) n=c(100,102,120,240) vr = value_remaining(x, n) hist(vr) best_arm = which.max(best_binomial_bandit(x, n)) # "potential value" remaining in the experiment potential_value = quantile(vr, 0.95) paste("Were still unsure about the CvR for the best arm (arm ", best_arm, "), but whatever it is, one of the other arms might beat it by as much as ", round(potential_value*100, 4), " percent.", sep="")