1
$\begingroup$

In simple words, how we can force any logistic regression coefficient be 1. Is there any steps/algorithm that used behind offset term?

$\endgroup$
  • $\begingroup$ I'm not really clear on what you're asking. $\endgroup$ – shadowtalker Apr 23 '16 at 19:06
  • $\begingroup$ I am confused that how we force coefficent to be 1. For example, in regression lm((y-x1)~1) it will force coeffcient of x1 be 1. How we use this trick in logisitic regression? $\endgroup$ – statuser Apr 23 '16 at 19:17
  • $\begingroup$ Yes I can add there, but what is algorithm or trick behind that? $\endgroup$ – statuser Apr 25 '16 at 5:00
3
$\begingroup$

There is no "trick" behind the use of an offset term ( * ), while the other coefficients will be estimated, that is, their values will be updated at each step of the IRLS ( * ) fitting algorithm, the offset coefficient will not be updated, it will keep its value of 1.

( * ) IRLS is iteratively reweighted least squares, the fitting algorithm usually used for generalized linear models such as logistic regression. For an overview, see Can you give a simple intuitive explanation of IRLS method to find the MLE of a GLM?

The offset term is included with a term offset(x1) in the model formula, or via the use of a separate offset= argument. The first way is the preferred one. This will force the coefficient of x1 to be one.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.