3
$\begingroup$

Suppose $Z_i$ are independent Bernoulli random variables with differing probabilities $P_i$. Also suppose weights $W_i$ are positive and constant.

Can you tell me the mean and variance for the random variable $S$ which is the summation of each weighted $Z_i$ (i.e. $W_iZ_i$). Furthermore is there a simple distribution for this random variable (similar to the Poisson Binomial distribution)?

Improve this question
$\endgroup$
1
  • 2
    $\begingroup$ What exactly are the weights supposed to mean? In what sense are you "summing" these variables? Are you literally forming the random variable $\sum_i w_i Z_i$ (as you write), or are you perhaps trying to form a mixture of their distributions, or maybe something else? And where you write that the weights are "constant" does that mean they are all equal to each other? $\endgroup$
    – whuber ♦
    Oct 10, 2014 at 16:26

1 Answer 1

Reset to default
5
$\begingroup$

I understand $Z_i$ takes value 1 with probability $P_i$ and $0$ otherwise.

Computing the mean of the weighted sum is simple: by linearity of the $\mathrm E$ operator,

$$ \mathrm E \sum_i W_i Z_i = \sum_i W_i \mathrm E Z_i = \sum_i W_i P_i. $$

As for the variance, since the variables are independent you can just sum their variances. So:

$$ \mathrm{Var} \sum_i W_i Z_i = \sum_i W_i^2 \mathrm{Var}Z_i = \sum_i W_i^2 P_i(1-P_i). $$

The distribution of $\sum_i W_i Z_i$ does not have a name, to my knowledge.

Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ To make it clearer that this answers (quite correctly!) one possible interpretation of the question, consider pointing out that the weighted sum of $n$ variables you are working with will not describe a distribution supported on the integers $\{0,1,\ldots,n\}$, which is the support of a Poisson-Binomial distribution (unless all the weights are equal to $1$). This contrast casts some doubt on whether your interpretation, albeit a natural one, truly is the intended one. $\endgroup$
    – whuber ♦
    Oct 10, 2014 at 20:33

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.