Suppose we wish to estimate the probability, PA, of some event A as outlined .As motivated by the result, suppose we repeat our experiment for a random number of trials, N. In particular, we run the experiment until we observe the event A exactly m times and then form the estimate of PA according to

Here, the random variable represents the number of trials until the m th occurrence of A.
(a) Find E [á¹”A]. Is this estimate unbiased?
(b) Would it be better to use á¹”A = m / N as an estimate?

Transcribed Image Text:

т- 1 PA N-1'