[R] reference for logistic regression

Robin Hankin r.hankin at noc.soton.ac.uk
Thu Oct 11 14:43:08 CEST 2007


> [snip]
>
> Whenever you use wikipedia you should be cautious of the quality of
> the information in the articles.  Generally the articles are good as a
> brief introduction but they can and do contain errors so you should
> check important facts and not take them at face value.  A person in
> one of my classes asked about the standard deviation and I suggested
> that they look at the wikipedia article on the topic.  Then I looked
> at it myself and saw that one of the things mentioned is that the
> standard deviation of the Cauchy distribution is undefined, which is
> true, but the reason given is because E[X] is undefined, which is not
> true.
>



The variance page has

<quote>
Many distributions, such as the Cauchy distribution, do not have a  
variance because the relevant integral diverges. In particular, if a  
distribution does not have an expected value, it does not have a  
variance either. The converse is not true: there are distributions  
for which the expected value exists, but the variance does not.
</quote>

which seems to be right.

I'd say that A => B but B =/> A together with A being true would be
expressed as "B is true because of A" which is pretty much what the  
standard deviation page
says.  Is this what you meant?



--
Robin Hankin
Uncertainty Analyst
National Oceanography Centre, Southampton
European Way, Southampton SO14 3ZH, UK
  tel  023-8059-7743



More information about the R-help mailing list