Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.



Re: Formula for Correlation Between Two Lognormal Variables
Posted:
Jun 8, 2012 9:15 AM


On 20120608 06:18:31 0300, aste said:
> Could somebody please post a link with the formulas and proofs? Even > when I use your exact keywords, my google results do not contain the > papers you are referring to.
<http://www.stuart.iit.edu/shared/shared_stuartfaculty/whitepapers/thomopoulos_some.pdf>
<http://stats.stackexchange.com/questions/6853/correlationoflognormalrandomvariables>
which includes a citation of above paper
Note that Google treats the search for "xxx yyy" (with the quotes) differently than just xxx yyy (without the quotes). A search for "lognormal correlation" (with quotes) was not very useful. Normally one tries several minor variations on the search as the exact search algorithm seems to be a carefully guarded secret that changes over time although the basic features are publicly documented. There is an arms race between the search algorithm revisers and the folks who game it for commercial gain as high rankings help sales for many things.
When quoted the string is expected to be matched as is while the unquoted version just finds the words in any order and at any spacing with the first being most important etc.
If you had a very large exhaustive statistical textbook the index under "formulae" will not usually be helpful for your question. Even the index for "correlation" is not going to be that helpful. But the index for "lognormal" is a good bet and you will be in good hands if there is a subindex entry for "correlation". Think of Google as an index and not as an artificial intelligence assistant. Google does not play Jeopardy even though the recent success of Watson from IBM Labs has gotten a fair bit of publicity.



