Talk:Poisson regression

Regularized Poisson regression

There is not a lot of literature on this, but I can't find anyting taht calls it "Regularized Poisson regression" although that seems a better name that Poisson Ridge Regression. Are there other names that are more popular?Billlion (talk) 11:28, 27 February 2019 (UTC)

Rating

I gave this a "Start-Class" rating on the quality scale. To improve in quality, this article needs at least the following:

1. A picture showing the conditional Poisson distribution as x changes.
2. A separate section on maximum likelihood estimation, giving the likelihood function with graphs of its surface, and going into greater detail on numerical methods for finding its maximum.
3. Discussion of the relationship to proportional hazard models.
4. A section on the use of Poisson regression in spatial models.
5. A section giving more examples in detail.
6. Better and deeper references.

This is a good topic, and it deserves a thorough treatment. —Aetheling (talk) 05:45, 26 June 2010 (UTC)

Ridge regularization

The present article states: "Regularization can be added to this optimization problem by instead maximizing

$\sum _{i=1}^{m}\log(p(y_{i};e^{\theta 'x}))+\lambda ||\theta ||^{2}$,"

But maximizing and a positive sign on the penalty term seem not to go hand in hand. Shouldn't it be a minus instead? — fnielsen (talk) 13:37, 6 April 2011 (UTC)

Yes, I believe you are correct. I've changed it. -BlueScreenD — Preceding unsigned comment added by BlueScreenD (talk • contribs) 03:51, 16 May 2012 (UTC)