H*****r 发帖数: 764 | 1 This isn’t over. This isn’t the one moment, the one stance, that lifts
the Ohio State football program -- the Ohio State University -- out of the
muck and mire.
There are two left, and they’re just around the corner: once athletic
director Gene Smith and president Gordon Gee are fired, the healing and the
rebuilding can begin.
Jim Tressel resigned as coach at Ohio State on Monday, and the first step
has been taken to end this 13-month ordeal of lies, denials and coverups
that threatens to bring ... 阅读全帖 |
|
a****m 发帖数: 693 | 2
model
主要是对他们原理的不同不是很理解:看了半天,大概有个认识了,把别人的slide上的
东西copy下
来,很是拗口。
总的理解是GEE 用的quasi-likelihood numerically solves their parameter, beta
and variance.
mixed model assume normal distribution, and analytically solve the
parameters using ML.
GEE vs MRM
• GEE not concerned with V (yi)
• GEE yields both robust and model-based std errors for beta ;
MRM, in common use, only provides model-based
• GEE solution for all kinds of outcomes; MRM needs to be
derived for each
• For non-no... 阅读全帖 |
|
TN 发帖数: 1870 | 3 【 以下文字转载自 Tennessee 讨论区 】
发信人: AOL (BeCozOfFirstName), 信区: Tennessee
标 题: Re: Gordon Gee to Return as Ohio State President
发信站: BBS 未名空间站 (Fri Jul 13 11:00:05 2007)
这事儿拖得有点久,突然发生还是有点诧异的。因为按理说他老婆抽大麻那个事情已经
比较早之前的事儿了。估计是内部矛盾长期不能调和,不过从Gordon Gee的功绩来看,
有的真的是能够真真切切感受到的,而不是在那里空放屁。
比如亚裔,少数民族在学校的比例,这个我感觉最深刻,几年时间发生了很大的变化。
他的campaign募捐做得很不错,医学院是芝麻开花节节高。
唯一缺憾的是在Gordon Gee在位期间,自从商学院的前任离职以后就没有找到合适的人
选,没有很好的人带领欧文走出困境。地域劣势不是一天两天了,到现在好像凸显得更
明显了。其实,还是管理的问题。因为法学院就排名没有下降。不过总体,Gee在位期
间确实做出了很多成绩,最起码,你有这么真切地喜欢过某位校长吗? |
|
TN 发帖数: 1870 | 4 去年九月Vanderbilt Employee Celebration Week,摆了一个Gee的假人供大家拍照,
我们挨个搂着假人照相。。。突然假人后面冒出一老头来。。。定睛一看。。。Gee本
人!!于是纷纷抛弃了假人,跟Gee老真人合影。九月份Nashville的天气又热又闷,不
过老人家还是很有耐心地一一满足了大家的要求。。。所以说挺nice的,起码面儿上平
易近人。 |
|
p**********l 发帖数: 1160 | 5 When we learned it in a consulting class, GEE was used to analyze correlated data consists of measurements taken on different clusters which are comprised of members or subunits, for example, longitudinal discrete data.
http://en.wikipedia.org/wiki/Generalized_estimating_equation
Generalized estimating equation
In statistics, a generalized estimating equation (GEE) is used to fit the
parameters of a generalized linear model where unknown correlation is
present.
The GEE allows for correlation wit |
|
s*********e 发帖数: 1051 | 6 just want to clarify your question.
are you specifically interested in the difference in estimation method
between gee and glmm?
or are you interested in the conceptual difference between marginal model
referred to gee and conditional model referred to glmm?
also, since the difference between gee and glmm is less pronounced in
gaussian models, I assume you are referring to non-gaussian models.
in any sense, this is an interesting topic. |
|
z******j 发帖数: 1265 | 7 the major disadvantage of GEE is that it only takes care of 1 hierarchical
structure and makes inference on the mean level, whereas HGLM can handle
several-layer structure and makes inference on the independent cluster level.
for example, there are 2 treatments for two groups of ppl. each person within
the group are observed several time points. In this case, the GEE model makes
inference on the effect of treatments. the GLM makes inferene on the person level. |
|
h*t 发帖数: 187 | 8 各位达人:
偶现在用R package gee 做一个GEE的regression,比如log(Y)=beta0+x*beta1,如果
我想测试H0:beta1=1,contrasts应该怎么写?
这个package里面没有给contrasts的例子,自己试了几次都不对。
另外,对同样的这组数据,如果用R做Generalized Linear Mixed-effected Model,用
哪个package的哪个命令比较好? |
|
h*****5 发帖数: 14 | 9 想建gee model,因为是特殊的分布,不能直接用proc genmod.
网上找了一个gee macro, 但总有一些错误。
有用过的大侠能否指点那里可以找到合适的code.
谢谢先。 |
|
j*****e 发帖数: 182 | 10 I have never used NLMIX macro. But I have used PROC NLMIX a lot. It is a
likelihood based approach, not GEE. You mentioned normal distribution. Are
you referrring to the random effectis treated normal? NLMIX can only
tolerate one random effect in the model. If you have multiple random effects
, you have to use GEE. Is NLMIX macro different from the PROC?
In general linear model, the random statement controls the design matrix and
the repeated statement control the residuals/sigma. In generalized |
|
h*****5 发帖数: 14 | 11 NLINMIX macro 可以fit gee model而proc nlmixed 不能。但NLINMIX macro只用于正
态分布,proc nlmixed 适用于更多的分布。
我已经proc nlmixed做了random effect zero-inflated model,现在想做gee model.找
不到现成的方法。 |
|
m*******7 发帖数: 244 | 12 Does any one have an electronic version of Mixed Effect Model, GEE class
notes ?
How is GEE related to Mixed Effect Model? Could anyone please explain to me?
Thanks. |
|
h****s 发帖数: 16779 | 13 Both GEE and Linear Mixed Effects Model (LMM) can be viewed as special cases
of the Generalized Linear Mixed Effects Model (GLMM):
GEE is the population averaged (PA) marginal model of GLMM, see Zeger, Liang
, Albert, 1988, Biometrics, 1049-1060.
LMM is the GLMM with normal distribution and identity link function, see SAS
help. |
|
a**a 发帖数: 8 | 14 Hi,
I am fitting a GEE model using gee R package, but I am not sure how to get
OR and its CI?
Thank you for your help, |
|
r****t 发帖数: 276 | 15 诸位大虾,在下试用GENMOD with repeated option去估计一个率,看起来很简单的
model,比如这样:
proc genmod data = dsn;
class subjid y x;
model y = x/link = logit distribution = binomial;
repeated subject = subjid/type=exch;
estimate 'rate a' x 1 0;
run;
可是一旦使用了repeated option 也就是GEE model,出来的率非常邪门。如果直接
proc freq; tables y*x得到的y=1 & x=1的率是70%左右,用GEE model求出来的居然在
30%左右?我注意到用exchangeable option 出来的working correlation〉0.95,这是
不是如此离奇的rate的原因亚?那位大虾对这种情况有经验望不吝赐教,谢先 |
|
r****t 发帖数: 276 | 16 窃以为把y编成1,2然后用default的order option,有用。赫赫
很同意“GEE的好处就是即使你指定的working correlation matrix是错的,最后的
mean model也不是很差”这句话,所以我才有这个问题。我用GEE有相当一阵子了,第
一次看到这种情况。我的数据是5个人review同一组病例给出的binary rating,怎么看
应该是clustered的数据。可这model QIC值是用independent structure来的最小,还
小那么多,拎不清哪
model |
|
r*****y 发帖数: 199 | 17 我想PharmD说的传统的survival就是把bmi都当成time independent variable,最多就
是adjust一下baseline BMI,你现在是要把它当成time dependent variable 来fit。
关于model fit可以去看看survival analysis using sas,2nd这本书,chapter5里面
有讲time-dependent covariates measured at regular intervals,第161页,你的
data manipulation可以在proc phreg里面完成,还有各种变化和讨论。
关于用survival还是longitudinal,我个人认为应该用survival,因为如果你把数据当
成longitudinal data来fit,其实你有monotone missing,因为一旦observe到event,
后面的纪录都删掉了,在MCAR的情况下,你可以fit GEE model,但你这个missing肯定
不是MAR了,所以结果应该会是biased。如果你有保留event发生后时... 阅读全帖 |
|
h******3 发帖数: 190 | 18 假设一个household里的人correlated.每个household的人数不一样这样的different
cluster sizes. 我用geeglm()试了下不行啊。
这种情况该怎么操作GEE。
用random effects model应该很容易,但是我需要用GEE。抓狂中。
多谢帮忙! |
|
|
b*******9 发帖数: 13548 | 20 【 以下文字转载自 Rite_aid 俱乐部 】
发信人: bactive79 (天南地北), 信区: Rite_aid
标 题: Gee~~~没笑死,就恶心死
发信站: BBS 未名空间站 (Mon Sep 12 21:49:00 2011, 美东)
原版 |
|
s*******a 发帖数: 705 | 21 Gordon Gee一去就把atheletic department解散了, 把AD也赶走了 |
|
TN 发帖数: 1870 | 22 肯定比在Vandy拿的少了不是一点
发信人: Asura (阿修罗), 信区: Tennessee
标 题: Re: Gordon Gee to Return as Ohio State President
发信站: BBS 未名空间站 (Thu Jul 12 21:26:39 2007)
He is probably going to take a pay cut. But it will still exceed 1 million
dollars.
the |
|
TN 发帖数: 1870 | 23 在Vandy当校长拿的就是很多,虽然Gee这次并不是为了钱去OSU,也很可能take a cut
但是大概还得维持这个数量级吧
其实Vandy哪点都不比OSU强,除了两方面:1)白妞pp指数;2)资金。。。 |
|
m********o 发帖数: 27 | 24 He makes $1.2m at Vandy. The reported basic salary that OSU will provide is
$775k, but I guess adding all the compensations & incentives, it wouldn't be
much less.
Vandy's president is definitely a better position than that of OSU's, but
7 years is probably enough for one's curiosity. Gee may have felt bored
there and decided to take OSU, a bulky public school, as his last toy.
cut |
|
TN 发帖数: 1870 | 25 【 以下文字转载自 Tennessee 讨论区 】
发信人: AOL (BeCozOfFirstName), 信区: Tennessee
标 题: Re: Gordon Gee to Return as Ohio State President
发信站: BBS 未名空间站 (Mon Jul 16 02:21:36 2007), 转信
Serious, I feel I got benifits. One example, you do not need to pay the
public
transportation fee, right? OK, another one, as a vanderbilt's faculty or
staff,
you were enrolled into the health plus program and you can have a better gym
facility. And, MRBIII finished construction at 2002, MRBIV is now under
construction now, only |
|
g********s 发帖数: 3652 | 26 今天'周四,Buck Gee和谭志明希望通过举办由亚裔科技团体参与的系列座谈和论坛,
与在硅谷企业担任管理阶层的华裔分享他们的经验,协助青年一代华裔主管在各大企业
中崭露
头角。'
请问这里说的‘系列座谈和论坛’时间地点是什么?
我们生死存亡的大事,请大家积极关注! |
|
r*******u 发帖数: 8732 | 27 【 以下文字转载自 HongKong 讨论区 】
发信人: rainbowwu (Ms.Wu :)), 信区: HongKong
标 题: hong kong bee gees:i started a joke
发信站: BBS 未名空间站 (Mon May 21 15:39:54 2012, 美东)
R.I.P. Robin Gibb. |
|
x*******1 发帖数: 28835 | 28 说自己SOS强, LSU和Auburn都要笑了。
说自己胜率高, 前面4只不败球队都要笑了。
说自己历史战绩好, LSU和鳄鱼都哈哈大笑。
这话怎么也不该gordon gee说阿。 |
|
b**j 发帖数: 20742 | 29 用你这个道理,你这句评论gordon gee的话似乎也轮不到一个utenn的扇子说吧 |
|
N*********6 发帖数: 4372 | 30 XiaoK,这个坑下午都被挖滥了,下面这么多牛鼻的比赛,随便挖个
好坑阿,你的三点Gee校长一个都没说阿 |
|
f*******u 发帖数: 2018 | 31 其实这种言论不会有law suit也不会损失什么,一帮人支持,一帮人反对,吵架,打口
水仗,结果创造很多gdp,大家high一把,人家在这个地位上也不用care这些,
brmj说这是political incorrect,其实不对,gee打死也不敢说老黑,女同志的坏话,
那可是立竿见影(参见mit校长的惨痛经历),对于这话,随便放炮,还能吸引眼球,
提高学校内的支持率,espn/bsu/tcu反对人根本就当不存在一样
这才是行为艺术的最高级别,mitbbs上行为艺术算个p呀,作人要行为艺术,就要行为
艺术到这种甚至更高的高度 |
|
N*********6 发帖数: 4372 | 32 看起来ESPN很开心啊,Gee终于意识到自己言论的不妥,要闭嘴了 |
|
a*****c 发帖数: 3525 | 33 NCAA是戏子。在戏子们看来,Gee是有身份的人,说话要负责任;而戏子是不负责,是
表演给人看的。 |
|
L*1 发帖数: 11537 | 34 He does not need to talk, just shut up.
As one commented on ESPN, Gee found nothing wrong with Trussel either. |
|
d*****i 发帖数: 1618 | 35 发财哥挺他,祝他安度晚年。Gee,一路走好!
6个伤心包,啥也不说了。go bucks! |
|
M****m 发帖数: 3243 | 36 乱挖学术坑,宗教坑的,迟早跟xiaok一个下场
发财哥挺他,祝他安度晚年。Gee,一路走好!6个伤心包,啥也不说了。go bucks! |
|
l*********e 发帖数: 680 | 37 11W今天早上还在说Gee不会有事的,不就是内部开玩笑被捅到外面了嘛。
唉 |
|
b********r 发帖数: 7725 | 38 正确打法其实就是我说的,没必要和他硬拼
基本思路是引他吸毒,吸毒越多次,减甲减血就越快
然后在不杀其他虫子的前提下在岩石背后躲起来,不用开枪
也就是说,至少要让他吸一次毒,不能杀了虫子而放任流毒不管
rich说得对,杀死任何一个虫子最后都能扩散成满屏毒泉
对于单人来说,我昨天晚上让他吸了三次,然后躲起来,差不多半小时他就挂了
然后他爆的时候我就死机了!!!
网上很多人确定Master Gee非常容易死机,多人就好点
所以今天我们要多人一起搞,死了没关系,只要有一个人躲就行
另外,石头后面好像只有一个人的空间,其他人很容易被k |
|
r*****2 发帖数: 1549 | 39 第一说下重复刷Master Gee的Bug
这个Boss 24小时才能杀一次,但是有一个Bug就是不用离线调时间也可以刷。 杀掉
Master后,如果不掉水晶,那个铁门就不会开,这时候退出再进来又可以杀了。
我今晚杀了7次Master (第7次掉水晶了), 3次和Leadingmore一起杀 (杀完死机一
次), 4次 Solo。 4次Solo都一次轻松搞定, 但是Co-op 失败率极高。 我来说说一
点心得吧。
1. 多人玩, 引怪比较混乱, 如果Master突然追第2人去了,毒就很容易散开, 所以
在Master破盾前, 其他人可以躲起来打酱油。 用一个枪狂来引和杀虫子就可以。 虫
子如果离Master太远,先别杀。 把Master引到虫子附近,然后再杀虫子,这样确保一
次毒就能被吸掉。 毒如果不能一次吸掉, 经常导致大混乱。
2. 如果不想打酱油, 那么你的人物一定要在Master的背后,那样不管怎么开枪,他都
不会追你。这样可以等他破盾后,在后面狂捅他菊花。 如果你在他前面,被他看到的
话,不开枪他也追你。
3. Master 的血越低攻击越高,半血后,经常能把我从满血打到剩1... 阅读全帖 |
|
j******n 发帖数: 21641 | 40 【 以下文字转载自 Sex 讨论区 】
发信人: beijingren (to thine own self be true), 信区: Sex
标 题: Re: Bee Gees - I Started A Joke
发信站: BBS 未名空间站 (Mon May 21 21:09:35 2012, 美东)
RIP |
|
l******m 发帖数: 31446 | 41 R.I.P.Robin Gibb... Stayin'alive forever...
STAYIN' ALIVE... Bee Gees |
|
l******m 发帖数: 31446 | 42 “你土鳖”上的留言。。。
“I finally died hearing you died. Words cannot express my sorrow. I keep
playing all bee gees songs and listened to them carefully and learn how deep
the words are. I have no idea why musicians of this caliber has to leave us
so soon. My heartfelt sorrow to his family and the only remaining brother.
Pl. give them strength to live on. Rest in Peace - you beautiful
soul!“
---Whytamin007 20 hours ago
“I grew up with your music and loved this song from day one and never
realized... 阅读全帖 |
|
L*********h 发帖数: 2617 | 43 谁给解释下bee gees为啥唱歌用那个调调?肯定很有讲头
有没有视频看一下他们平时说话什么声音? |
|
b******n 发帖数: 54 | 44 Just start learning how to use PROC GENMOD, :-), Following are quote from SAS
manual:
The analysis of correlated data arising from repeated measurements when the
measurements are assumed to be multivariate normal has been studied
extensively. However, the normality assumption may not always be reasonable;
for example,different methodology must be used in the data analysis when the
responses are discrete and correlated. Generalized Estimating Equations (GEEs)
provide a practical method with reaso |
|
h*****5 发帖数: 14 | 45 谢谢!
貌似GLIMMIX只能做MIXED EFFECT MODEL USING RANDOM STATEMENT,没有REPEATED
STATEMENT.
想找一个GEE macro,然后改写CODE, 插入一个TRUNCATED POISSON 分布。能否用SAS
实现? |
|
h*****5 发帖数: 14 | 46 谢谢大家。刚看了下,nlinmix macro只能用于正态分布。正在看proc GLIMMIX 和
macro,希望能work。
to hypnos: 太菜了,自己写不出gee code. 方便的话,能否借鉴一下?非常感谢。 |
|
b*****e 发帖数: 223 | 47 GEE 不熟悉啊,学 LME 的时候也没听说啊
下面是两个学统计的网站,随便搜搜还有不少比较好的网上教材
statnote
softstat
me? |
|
m*******7 发帖数: 244 | 48 What are the differences between random effect model and GEE? |
|
e***o 发帖数: 180 | 49 this makes sense for Proc Mixed. thanks
can somebody use GEE for clustered data, say using Proc Genmod, if the
response is not continuous? |
|
a****m 发帖数: 693 | 50 感觉GEE就是GLM的延伸,引入了correlation matrix for responsive variable, 用来
计算
covariance matrix, 但是多了个 over-dispersion parameter.
mixed models, 引入了random effects, 原来的correlation 来自于random effects.
不知道两种方法的优势各在哪里? |
|