广义 R 平方 (Naglekerkes R^2)
does anybody know how to compute Naglekerkes generalized R Squared for GLMs using R? And does it makes any sence to use it for count data regression?
References:
"Scroll down at to the bottom of this article!"
如果你对这篇内容有疑问,欢迎到本站社区发帖提问 参与讨论,获取更多帮助,或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱,如果其他用户或者作者回复了您的评论,将不能在第一时间通知您!
发布评论
评论(1)
尝试
在各种包中查找函数:最有用的可能是
rms
包(Design
包的更新版本)中的lrm
和
。尽管如果“计数数据回归”指的是泊松模型而不是二项式/二元模型,则fmsb
包中的 >NagelkerkeR2lrm
不会为您做到这一点。但另一个应该。我不明白为什么它不适用于计数数据(泊松回归)......尽管总是有人争论 R2 的各种类似物一旦超出了 R2 的领域有多大用处普通线性模型...
Try
This finds functions in various packages: probably most useful are
lrm
in therms
package (an updated version of theDesign
package) andNagelkerkeR2
in thefmsb
package. Although if by "count data regression" you mean a Poisson rather than a binomial/binary model,lrm
won't do it for you. But the other should.I don't see why it wouldn't apply to count data (Poisson regression) ... although there is always debate about how useful the various analogues of R2 are once you get outside the realm of ordinary linear models ...