0, where B(a, B) = r(a)r(3)/r(a+3) is the beta function. The marginal distribu- tion of j purchases of Brand 1 on k trials is obtained by compounding the bi- nomial and beta distributions. The unconditional probability of j purchases of Brand 1 on k trials is given by P(j;k, a, 6)
...= ($) ^^^^T[!tIT^ ' 3 = o, U.. -X (9) This is the beta-binomial distribution whose mean and variance are given by and UAR Til - M(a+B+k) and VAR LjJ - {a+3)^(a+B+l ) • The expression for the unconditional probability of j successes on k trials in equation (9) can also be written in terms of the transformed parameters y and ({) by making the substitutions for a and B as follows: , = lilMi, and 6 =(1:m1(M) (10) Parameter Estimation Given the purchase frequency data (i. E. , N-'s) the likelihood function J can be written as where the term P.
What to read after Sample Size Requirements for Zero Order Models?
You can find similar books in the "Read Also" column, or choose other free books by Manohar U Kalwani to read online
User Reviews: