+ Reply to Thread
Results 1 to 2 of 2

Thread: Generating correlated random vars with Cholesky

  1. #1
    Points: 225, Level: 4
    Level completed: 50%, Points required for next Level: 25

    Posts
    1
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Generating correlated random vars with Cholesky




    Hi everyone,

    Can somebody please explain to me what is the intuition behind the cholesky decomposition when generating correlated random variables? The funny thing is that i know how to calculate it, the thing is that I have no idea why it works.......

    In my Monte Carlo simulator i generate a choleski deco on correlation matrix, mulitply it with vector of uncorelated standard normals , rescale the normals and get nice results...But why does multiplying Cholesky lower triangular by uncorelated vector gives me correlated one? I saw some simbolic, very general explanation but somehow I just dont get it.......

    Many thanks

    Jerzy

  2. #2
    Super Moderator
    Points: 10,251, Level: 67
    Level completed: 51%, Points required for next Level: 199
    Dragan's Avatar
    Location
    Illinois, US
    Posts
    1,841
    Thanks
    0
    Thanked 158 Times in 141 Posts

    Re: Generating correlated random vars with Cholesky


    Quote Originally Posted by panjerzy View Post
    Hi everyone,

    Can somebody please explain to me what is the intuition behind the cholesky decomposition when generating correlated random variables? The funny thing is that i know how to calculate it, the thing is that I have no idea why it works.......

    In my Monte Carlo simulator i generate a choleski deco on correlation matrix, mulitply it with vector of uncorelated standard normals , rescale the normals and get nice results...But why does multiplying Cholesky lower triangular by uncorelated vector gives me correlated one? I saw some simbolic, very general explanation but somehow I just dont get it.......

    Many thanks

    Jerzy
    Well, start with two variables. That is, given that X and E are iid standard normal, show that Y has correlation with X of r.

    Y=rX+\sqrt{1-r^{2}}E

+ Reply to Thread

           




Tags for this 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