求助：proc nlmixed 的parms依据什么设置？

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

最近正在学proc nlmixed 处理panel data，遇到下述例子，请问高手：parms的设置有什么依据，或者说怎样设置？

1.

options ls=80 ps=60;data one; infile 'isjall.dat';
input country gender bigvtlim jp average unemrate westeur
postcom centeur easteur medvoc seceduc thdeduc;

proc nlmixed;
parms gamma00=2.5 gamma01=0.5 gamma10=0.5 gamma11=0.1
su02=1 su12=1 cu12=0.5 beta3=0.5 beta4=0.5;
beta0=gamma00+gamma01*unemrate+u0;
beta1=gamma10+gamma11*unemrate+u1;
eta=beta0+beta1*gender+beta2*medvoc+beta3*seceduc+beta4*thdeduc;
expeta=exp(eta);
p=expeta/(1+expeta);
model bigvtlim ~ binary(p);
random u0 u1 ~ normal([0,0],[su02,cu12,su12]) subject=country;
run;

2

proc nlmixed data=foxngeese1 maxiter=1000;
  title2 'Model A: Unconditional logistic growth trajectory';
  parms G00=12 G10=.1 s2e=20 s2u0=.3 cu10=0 s2u1=.01;
  num = 19;
  den = 1 + G00*(exp(-(G10*GAME +u0 +u1*GAME)));
  model NMOVES ~ normal(1+num/den,s2e);
  random u0 u1 ~ normal([0,0],[s2u0,cu10,s2u1]) subject=id;
run;

3

proc nlmixed data=foxngeese1 maxiter=1000;
  title2 'Model B: Effect of READing scores on 'intercept' and 'slope'';
  parms G00=12 G10=.12 G01=-.4 G11=.04 s2e=14 s2u0=.5 cu10=0 s2u1=.01;
  num = 19;
  den = 1 + G00*(exp(-(G01*READ + G10*GAME +G11*READGAME +u0 +u1*GAME)));
  model NMOVES ~ normal(1+num/den,s2e);
  random u0 u1 ~ normal([0,0],[s2u0,cu10,s2u1]) subject=id;
run;

多谢救危！！！！！

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏0 回帖

关键词：proc nlmixed nlmixed parms Mixed Arms proc 设置 nlmixed parms