j**r
发帖数: 68 | |
X*********e
发帖数: 253 | 2 glm(........ family=binomial(logit))
【在 j**r 的大作中提到】
: thx
|
p*******t
发帖数: 501 | 3 如果有bayesian statistic的话，R是不是只能用MCMC来simulate?
【在 X*********e 的大作中提到】
: glm(........ family=binomial(logit))
|
X*********e
发帖数: 253 | 4 no, its depend on what is your posterior distribution.
【在 p*******t 的大作中提到】
: 如果有bayesian statistic的话，R是不是只能用MCMC来simulate?
|
p*******t
发帖数: 501 | 5 就是binary choice logit model
顺便问，这个跟probit model的结果有本质性的区别么？
【在 X*********e 的大作中提到】
: no, its depend on what is your posterior distribution.
|
X*********e
发帖数: 253 | 6 hehe, sorry,i don't know
knew very few about probit model
【在 p*******t 的大作中提到】
: 就是binary choice logit model
: 顺便问，这个跟probit model的结果有本质性的区别么？
|
x********4
发帖数: 405 | 7 logit assumes the error term follows log normal distribution which cannot be
negative, while probit assumes the error term follows normal distribution
that can be negative......
【在 p*******t 的大作中提到】
: 就是binary choice logit model
: 顺便问，这个跟probit model的结果有本质性的区别么？
|
p*******t
发帖数: 501 | 8 我知道.......所以我想问又没有本质的区别.....比如substitution pattern之类的
be
【在 x********4 的大作中提到】
: logit assumes the error term follows log normal distribution which cannot be
: negative, while probit assumes the error term follows normal distribution
: that can be negative......
|
j**r
发帖数: 68 | 9 gotcha!
thanks
【在 X*********e 的大作中提到】
: hehe, sorry,i don't know
: knew very few about probit model
|
t****g
发帖数: 715 | 10 Wrong. Of course logit error term can be negative. We assume that the
distribution of error term in logit model is logistic.
be
【在 x********4 的大作中提到】
: logit assumes the error term follows log normal distribution which cannot be
: negative, while probit assumes the error term follows normal distribution
: that can be negative......
|
t****g
发帖数: 715 | 11 No big difference between logit and probit, however, you choose one of them
based on your preference. For instance, it may make your life easier to
choose probit, as its error is normal, which can either fit your model
better or make your presentation more reasonable for people outside
econometrics(you know, sometimes it is hard to explain to dumb guys why you
need a logistic distribution instead of normal). Historically, logit was
thought easier to compute during the period when computing distr
【在 p*******t 的大作中提到】
: 我知道.......所以我想问又没有本质的区别.....比如substitution pattern之类的
:
: be
|
p*******t
发帖数: 501 | 12 知道了，谢谢啦
them
you
,
【在 t****g 的大作中提到】
: No big difference between logit and probit, however, you choose one of them
: based on your preference. For instance, it may make your life easier to
: choose probit, as its error is normal, which can either fit your model
: better or make your presentation more reasonable for people outside
: econometrics(you know, sometimes it is hard to explain to dumb guys why you
: need a logistic distribution instead of normal). Historically, logit was
: thought easier to compute during the period when computing distr
|
f*******r
发帖数: 257 | 13 Generally there are two ways of deriving the models for binary data:
1. latent variable approach: y*=x \beta + u. Here y* is a continuous
latent variable. then u can be treated as logistically or normally
distributed, which corresponds to logit or probit model. This approach is
modeling the underline variable behind the binary variable y.
2. link function approach. This approach models the binary variable y
directly: y=g(X \beta), where g() is the inverse link function. Basically
it maps |
