Given a biased die with d faces, you are given results of n die rolls. I need to calculate the confidence intervals of the probabilities of each of the d outcomes of the die.
A solution in R! - even better.
Asked
Active
Viewed 38 times
0
-
1If you just need individual confidence intervals, your question is answered at https://stats.stackexchange.com/questions/4756. Are you therefore implying you need a simultaneous confidence region? – whuber Mar 26 '20 at 18:42
-
don't think I know enough to follow that solution. is there a simple R implementation as a function of d and n? – Eric Mar 26 '20 at 19:31
-
Simple (one-line) `R` implementations are provided in answers in that thread. – whuber Mar 26 '20 at 19:33
-
sorry for the novice follow-up question: the answer is for fail/success Bernoulli trials. so, if I want to the CI for the probability of outcome "5" for a 5-faced die. I would classify "1" through "4" as "fail" and "5" as "success" and then apply the methodology in the answer? – Eric Mar 26 '20 at 19:46
-
That's exactly right. You would need to repeat that procedure for each possible value of the die. – whuber Mar 26 '20 at 19:47