+ Reply to Thread
Results 1 to 3 of 3

Thread: Combining two normal distrubtions, 1 of which is for a project which may not happen

  1. #1
    Points: 220, Level: 4
    Level completed: 40%, Points required for next Level: 30

    Posts
    5
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Combining two normal distrubtions, 1 of which is for a project which may not happen




    Hello,

    I have to 2 projects:
    Project A: Median: $450k, Variance: 625
    Project B: Median: $750k, Variance: 1250
    Combined Freqency Distribution: Median: $1200k, Variance: 1875

    Please could you tell me how to find the Combined Frequency Distribution (for both projects combined) if Project B only has a 60% likelihood of happening?

    Thank you for your help.

    Mr E Guest

  2. #2
    Points: 220, Level: 4
    Level completed: 40%, Points required for next Level: 30

    Posts
    5
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Combining two normal distrubtions, 1 of which is for a project which may not happ

    sorry I'm guessing from the no replies that the question doesn't really make sense? or is this a very hard problem to solve? Should I approve the problem in a different way?

    Thanks,

    Mr E

  3. #3
    TS Contributor
    Points: 22,410, Level: 93
    Level completed: 6%, Points required for next Level: 940

    Posts
    3,020
    Thanks
    12
    Thanked 565 Times in 537 Posts

    Re: Combining two normal distrubtions, 1 of which is for a project which may not happ


    Your question does make sense, and I hope you understand that there are quite a number of new posts in TS everyday so we may miss your thread as it sink beyond the most updated 10.

    So I assume that in the original setting, you are having two random variables A, B (representing the two projects) and you have some results about the sum of them

    C = A + B

    (which you called combined of them). The result itself is always valid if they are independent and you are talking about mean instead of median.

    The situation you mentioned I guess means the following mixture:

    C = \left\{\begin{matrix} A + B, & \text{with probability 0.6} \\
A, & \text{with probability 0.4} \end{matrix} \right.


    The distribution of this mixture can always be found (btw finding a distribution in probability context means find its cumulative distribution function), you just need to provide the distribution of each of them. If you are only interested in the mean and variance, you may apply the law of total variance to calculate it, by rewriting

    C = (A + B)I + A(1 - I) = A + BI

    where

    I \sim \text{Bernoulli}(0.6) which is independent of A, B.

    So in this case

    E[C] = E[A + BI] = E[A] + E[B]E[I] = E[A] + 0.6E[B]

    Var[C] = Var[A + BI]

    = Var[A] + Var[BI]

    = Var[A] + Var[E[BI|I]] + E[Var[BI|I]]

    = Var[A] + Var[E[B]I] + E[Var[B]I^2]

    = Var[A] + E[B]^2Var[I] + Var[B]E[I^2]

    = Var[A] + E[B]^2 \times 0.6 \times 0.4 + Var[B] \times 0.6

    So you can just plug in the values given to obtain the result.

+ Reply to Thread

           




Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats