Re: logistic regression

From: Richard Ulrich (Rich.Ulrich_at_comcast.net)
Date: 01/21/05 

Date: Fri, 21 Jan 2005 11:36:59 -0500

 - here's a wordier answer -

On Thu, 20 Jan 2005 22:34:27 -0600, "raed shatnawi"
<rshatnaw@cs.uah.edu> wrote:
 
> Hi all,
>
> what is the best book that explains multivariate statistics?

Where are you starting?
When you browse in a dozen stat textbooks in the
library, which ones seemed to make the most sense?
 - You might find something good by browsing, but,
too often, the best books are already on-loan. IF you
say what seems good, someone might say what seems
similar, but better.

>
> can anyone explain to me the difference between least sqaure regression and
> logistic regression?

The "logit" is defined as log (p/q) where q=(1-p), and
p is the probability for a dichotomous variable which
has been observed or estimated for a group.
[You might look up "probit" for comparison.]

You can use Y= < 0 or 1 > as criterion in Ordinary Least Squares.

There's an obvious problem in trying to replace that with
a simple transformation for an individual, rather than
a group, as log(p/q) -- "Logistic regression" (LR) has easy
solutions (like OLS) when the data are expressed as groups
where p is not 0 or 1. And logistic regression requires
a more subtle solution, when the data are presented
as individuals whose scores are 0 or 1, since you can't
take the log of 0 or infinite.

Practically speaking, there is often very little difference
in tests or prediction between OLS regression on 0/1
outcome, and LR by maximum likelihood methods.
LR usually represents the more appropriate theoretical
model So far as I know, the computer programs for Logistic
regression all remain less effective at warnings and
diagnostics than what you can get on OLS regression.

http://www.pitt.edu/~wpilib/index.html

Relevant Pages

  • Re: Single-Factor-Cox-Regression
    ... allowed within Cox-Regression) ... Like the logistic regression, which it is sort-of an extension ... Cox regression models hazards and hazard ratios. ... Logistic regression is to require at least 20 more cases in the ...
    (sci.stat.math)
  • Re: Approximate solution to linear regression
    ... Construct an ensemble of regression models, ... these are binary variables. ... So my idea to use a logistic regression to classify 15% of the ...
    (sci.stat.consult)
  • Re: Adjusting
    ... some about unequal sample sizes. ... >> One way is to just do weighted regression, ... > iterations of the logistic regression fitting process. ...
    (sci.stat.math)
  • Re: Logistic regression or Poisson regression (log linear)
    ... I've been looking at some analyses with similarly sparse data. ... prior to logistic regression will probably lose a lot of information. ... > treatment more often than with the experimental treatment. ...
    (sci.stat.consult)
  • Re: Regression Inference and Data Splitting
    ... The statement is: " Regression ... is logistic regression, it has relevance to all regression-type models. ... Internal validation of predictive models: ... Several internal validation methods are available that aim to ...
    (sci.stat.math)