b****k
发帖数: 409 | 1 因人而异，俺是mix random timed做的。挺有估分价值的。第一遍70%，很有希望250+
。错题是下面复习的重点。第二遍也可以分章节做 |
|
j*****9
发帖数: 193 | 2 我买了6个月的题库。打算第一遍按照章节来做题，把每章节知识点吃透，然后进入下
一章。待全部做完题库后，就reset题库，按照timed mixed模式做自测，看看还有哪些
知识点没有记住。 |
|
l*****1
发帖数: 137 | 3 不知道哪位做过GaAs/AlGaAs多量子阱的degenerate four wave mixing实验？
我想知道需要什么样的detector去测量衍射信号？一般衍射效率能达到多少？对激光波
长是否非常敏感？有可能用IR viewer或者IR card之类的东西直接看到衍射光吗？
我们现在的实验条件如下：
250Khz， 200fs， 两束泵浦光10mW，探测光1mW，温度80K，激光波长785nm。 |
|
w******a
发帖数: 27 | 4 要解MILP ( mixed integer linear programming)问题，请问Matlab 有没有解这类问
题的ｔｏｏｌｂｏｘ, 或者第三方提供的免费软件包。 |
|
|
d***n
发帖数: 40 | 6 Hi all,
The University of Auckland is looking to find a suitable candidate to
conduct trans-disciplinary research in the field of coastal
management. Please forward this to all who may be
interested/suitable. Contact details and instructions on how to apply
are below.
PhD scholarship, coastal management
University of Auckland, New Zealand
A mixed modelling collaborative framework for coastal management in
Auckland, New Zealand
NIWA’s Coastal OBI framework offers funding for a PhD project aimed |
|
|
t*a
发帖数: 117 | 8 The January SIGSTAT meeting will be held on Wednesday, January 21, from
12:30 to 1:30 in Room S3031, 1800 M St, NW (directions can be found on
the website below).
The topic is "PROC MIXED Part 3: Model Development and
Interpretation".
Part 1 applied EDA techniques to help in specifying an initial model.
Part 2 examined the issue of selecting an appropriate covariance
structure using the variogram as the main tool.
Part 3 will cover the specification of the mean part of the model,
plotting intera |
|
R****9
发帖数: 815 | 9 Thanks!!!
糊涂的地方是：
哪几个该放class里: subid, time, eye?
哪几个该放random里: intercept, subid*eye?
model 1里有random，为什么没有repeated? class里为什么没有time?
model 4 可行吗
4. proc mixed;
class eye time;
model score = time eye time*eye/solution;
repeated time /subject=subid;
run; |
|
W**********E
发帖数: 242 | 10
我觉得你可以考虑你的模型３
１。 人－随机１５０个
２。 人的眼睛（３个因素：左，右，双）固定
３。 每个位置５个时间点。固定
１＞２＞３
类似于一个SPLIT－SPLIT PLOT 设计。 人分３个位置，每个位置又分５个点。但你可
以用其他COVARIANCE STRUCTURE 来模拟WITHIN EYE POSITION的COVARIANCE。
3. proc mixed;
class eye time subid;
model score=eye time time*eye/ddfm=kr;
***subid*eye=error term for testing H0: eye=0
random subid subid*eye;
***subject=subid*eye is the unit within which time factor is nested;
repeated time/ type=ar(1) subject=subid*eye;
run; |
|
b*******g
发帖数: 513 | 11 google "proc mixed equal variance", you can find something at least related. |
|
d******3
发帖数: 93 | 12 In my generalized linear mixed model with logit link, why is DF=0 in the
Solution of Fixed Effects for a categorical variable?
I have checked freqencies and it >35 for each cell. |
|
a********a
发帖数: 346 | 13 不明白在mixed model 中random effect and repeated 的区别，请高人指点一下。 |
|
D******n
发帖数: 2836 | 14 They control different things, should check the mathematical model that
MIXED uses from the SAS documentation, it is very clear. But, as far as my
experience is concerned it is not very straightforward to see what effect
will be generated by using certain parameter combination of random and
repeat, u have to "figure" it out. |
|
q*******t
发帖数: 18 | 15 The residual plot showed non-normality. Will the inference from Proc mixed
still O.K to use?
Does "diff=sattherth" can handle unequal variance?
Thanks |
|
h******e
发帖数: 1791 | 16 这段code：
proc mixed ..........
class a；
model...........
random ......../ subject = a.......
中的"subject = a"是什么意思呀？在model中起到什么作用呀？谢谢。 |
|
f*********e
发帖数: 45 | 17 Usually, when fitting mixed model , if we run across the un converge problem
,
what we should do ? what I can think about is
1.try different type of variance structure like CS ...
2, specifiy different estiamation method. (mle - reml)
3. reduce covariates.
What do you think about it ?
Thank you so much! |
|
f*****a
发帖数: 156 | 18 这段SAS code转成R应该怎么写？
proc mixed data=temp;
by x1 x2 x3;
class x4 x5 x6 x7 x8;
model y_&&t&i = /s;
random x4*x5 x6 x7(x6) x4 x6*x4 x8 /s;
ods output solutionf=AAA solutionr=BBB covparms=cov; run;
run;
试着写了个，好像不行，不一样：
fit1=lmer(y~1+(1|x4:x5)+(1|x6)+(x6/x7)+(1|x4)+(1|x6:x4)+(1|x8),data=temp)
summary(fit1)
fit2=lme(y~1,random=list(~1|x4:x5,~1|x6,~x6/x7,~1|x4,~1|x6:x4,~1|x8),data=
temp)
summary(fit2) |
|
s*********t
发帖数: 3 | 19 有个关于Generalized Linear Mixed Models(GLMMs）的问题：
请问在R或SAS中能否得到 covariance matrix between fixed effect estimators (\
beta) and variance component estimator for random effect(\sigma)
在R或SAS中可以得到fixed effect estimators的variance matrix，也可以得到
variance component estimator(\sigma)的variance。但是能否输出它们两者的
covariance呢？
多谢～ |
|
s*********t
发帖数: 3 | 20 看了proc mixed的manual, 没有相关输出项。proc glimmix倒没看。
我目前的做法是在R中写出loglikelihood function，然后用nlm,可以得到Hessian
matrix，再求逆；
多谢大家的回复～ |
|
s*r
发帖数: 2757 | 21 proc glm 和 proc mixed有不同的要求 |
|
a********a
发帖数: 346 | 22 Is there a way to get residual from mixed model just like from a linear
model to have a qq plot to see the normality of the data?
Thanks |
|
a********a
发帖数: 346 | 23 I am confused for the following two statements when running mixed model.
random id;
random intercept/sub=id;
I thought they are same since both are treating id as random. But SAS gives
different results. So they should not be same. Can somebody tell the
difference? Thanks |
|
k*****t
发帖数: 375 | 24 如题了。对 proc mixed 非常不熟悉。
谢啦！ |
|
h******e
发帖数: 1791 | 25 用proc mixed分析repeated measures，一个fixed effect的不同level的variance可以
在covariance matrix table里找到。然后,能不能估算两个不同level的variance的比
值的confidence interval？ |
|
s********9
发帖数: 74 | 26 If I wanted to check the effect of A*B in proc mixed;
A: 1 2 3
B: 1 2 3 4
What could I do to make the group A=3 and B=4 as reference?
When I used estimate, there are lots of lines of A*B showed up as 0......
I want to try contrast statement, but it looks very confusion!
Anyone has experience? |
|
b*******a
发帖数: 7 | 27 有没有谁知道general linear mixed model有哪些limitation呢?
google了一圈都没看到有写GLMM的drawbacks的
谢谢 |
|
e***o
发帖数: 180 | 28 this makes sense for Proc Mixed. thanks
can somebody use GEE for clustered data, say using Proc Genmod, if the
response is not continuous? |
|
W****N
发帖数: 43 | 29 请问proc mix 在生统中具体的应用有哪些啊？和proc glm有啥区别？谢谢指教！ |
|
f**********t
发帖数: 1001 | 30 好弱的问题，搞了我一晚上。
Proc mixed
class gender nationality;
model salary = gender nationality gender*nationality;
run
其实我就是想看几个independent variable的coefficient。找不到。也不能
像Proc reg那样用plot来看到。在网上一顿搜资料也发现讲得好复杂。该怎么弄
呢？感觉这个应该很基本吧。太感谢了！ |
|
d*******1
发帖数: 854 | 31 Proc mixed
class gender nationality;
model salary = gender nationality gender*nationality/solution;
run; |
|
a****m
发帖数: 693 | 32 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？ |
|
s*r
发帖数: 2757 | 33 gee似乎主要还是在categorical outcome上吧，proc genmod/logistic就可以实现
但在clinical trial里，repeated measure categorical outcome似乎还是用glimmix
多，glimmix就是反复调用proc mixed |
|
a****m
发帖数: 693 | 34
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 |
|
P****D
发帖数: 11146 | 35 http://rapidshare.com/files/34645943/
SAS.Publishing.SAS.for.Mixed.Models.
2nd.Edition.Mar.20062007-06-01-234545.zip
Make the above three lines into one link. It seems like mitbbs doesn't
display the whole link if it is very long. |
|
o*****e
发帖数: 50 | 36 I think random effects should not appear in model statement for proc
mixed. |
|
k****i
发帖数: 347 | 37 不用吧，他的objective已经说的很明确了
OBJECTIVE: 比较每个时间点TREATMENT EFFECT ON 听力
model我觉得没什么问题。我的问题是为什么用glimmix而不是mixed
你的听力score不是normal distribution么？ |
|
k****i
发帖数: 347 | 38 这个normal是针对residual，也就是误差项的假设，所以和有几个病人没有太大关系。
只要你的听力score不是太极端的分布（只有1,2,3那种），每次测量的误差大概是
normally distributed就可以了
我假设你的问题是，考虑到within-patient的repeated measures之后，总体的
treatment有没有效果。下面这个思路有点像paired t-test
假设delta=right-left.
proc mixed data=..;
CLASS id month;
MODEL delta = /SOLUTION;
repeated month / sub=id type=ar(1);
RUN;
intercept就是estimated difference between treatment and control, after
accounting for within-patient repeated measures
有可能model里要额外adjust for左耳的score和month。当然细究还有把ar(1)换成
spatia... 阅读全帖 |
|
w***o
发帖数: 1411 | 39 我没理解你的modle，我想比较的是：1.确定时间点，比较treatment effect. 2.
确定组（treatment or control),比较听力的时间趋势，即各个时间点的变化。
我对mixed model理解不深，请教你一下，repeated statement 和random statement,
区别。
我以为这个模型，不用repeated 而用random.
p |
|
k****i
发帖数: 347 | 40 上一个回复里，只对control group的code应该在where里加上group=0
整体时间上的趋势，可以假设时间是continuous的。
proc mixed;
class id;
model y=group month group*month/s;
random int /sub=id;
run;
如果interaction不significant,就去掉
square |
|
d****z
发帖数: 301 | 41 这些天一直折腾mixed model,一直报错infinite likelihood,而且数据是over time的,
我想用repeated procedure,但是死也没有看明白到底怎么用.有高人能指点下么... |
|
z****n
发帖数: 67 | 42 请问SAS中proc mixed为什么没有Rsquare的输出？ 理论上可以自及计算吗？
如果没有Rsquare怎么判断模型好坏呢？ |
|
p*******r
发帖数: 1951 | 43 首先要明确你要做什么比较？看你说的意思你的null hypothesis是不是
H0: treatment effect = time2 - time1 = 0 ?
如果是，那你的dependent variable 完全可以用 effect = time2 - time1。而不是单
独的time1, time2。
其次，你的数据显示不同的subject有完全不同的时间点， 看你的SAS程序里面也根本
没有用到时间变量week。所以你想做的不是longitudinal data analysis？你只是想比
较不同treatment, pre-, post- response的差别。这样的话用repeated measure
ANOVA就成了。
很久没有用过mixed model了。说的不对的地方大家指教。 |
|
c*********r
发帖数: 397 | 44 we were free to use it that time. he had phd of psychiatry instead of
statistics, but achieved a great success in biostatistics.
he is very smart , humorous and talented, plays guilta well and sings
the songs he created himself.
what I emphasized here is not I mentioned above, instead, is the course of
longitudinal analysis with mixed model methodology and application. |
|
|
w**********y
发帖数: 1691 | 46 3. SAS for mixed models. Helps me coding
I bet this one might be a good reference and fits your work requirements
. |
|
P****D
发帖数: 11146 | 47 Linear Mixed Models for Longitudinal Data by Verbeke and Molenberghs
tons of SAS examples
I think all Littell's books are pretty good. He is really good at teaching. |
|
j******a
发帖数: 194 | 48 请教大家一个问题，我要做一个mixed linear model
y_ij= beta_0+beta_1*x_ij+S_j,
其中 x_ij是一个fixed effect,每个y_ij都对应一个值，而S_j有4个值，不同的y可能
对应不同的值。如果写成matrix 形式，
Y= X b + Z S,
Z是由 (0,...,1,0)组成的indicator矩阵，但是问题在于有5%的数据中s_j的值未知。
请问如果不直接删除这些数据，有没有什么好的方法可以定义S或者Z ？或者简单来说
，怎么处理这样的missing covariates最好？谢谢。
我想用Bayesian的方法做，但如果能提供一些frequentist的方法做参考，也非常感谢 |
|
a********s
发帖数: 188 | 49 Does anyone know how to handle the situation that the covariance structure
for a mixed model depends on time? |
|
b*****e
发帖数: 223 | 50 AR?
Proc Mixed has the option to define AR covariance structure |
|
