question on multinomial logistic regression
- From: "godspeed.wang@xxxxxxxxx" <godspeed.wang@xxxxxxxxx>
- Date: Tue, 18 Mar 2008 07:57:33 -0700 (PDT)
Dear All:
I'm currently trying to do classification via multinomial logistic
regression. However the Hessian of the log-likelihood function is more
complex/large than that in binomial case. Suppose I have to divide
data into K groups. Can I do/train k binomial logistic regressions for
each group separately, and then use Bayesian formula to estimate p(y=k|
x).
.
- Prev by Date: Re: what can you conclude about the validity of the system (help)
- Next by Date: Can any one help me calculate a statistical probability
- Previous by thread: what can you conclude about the validity of the system (help)
- Next by thread: Can any one help me calculate a statistical probability
- Index(es):
