[R-sig-ME] Convert SAS codes into R
Muhammad Yaseen
myaseen208 at gmail.com
Sat Oct 23 05:36:39 CEST 2010
Hi Dear,
I'd highly appreciate if someone can help me to convert these SAS
codes into R. Thanks
/*-------------------------------------------------------*/
/*---Data Set 11.5, Uniformity Trial ---*/
data spatvar;
input rep bloc row col yield;
datalines;
1 4 1 1 10.5411
1 4 1 2 8.5806
1 2 1 3 11.2790
..........................
..........................
..........................
4 15 7 8 10.4298
4 14 8 5 7.3220
4 14 8 6 10.5104
4 15 8 7 12.6808
4 15 8 8 10.4482
;
/*-------------------------------------------------------*/
/*-------------------------------------------------------*/
/*---Output 11.1 ---*/
title 'no nugget - spherical covariance model';
proc mixed data=spatvar;
model yield = ;
repeated / subject = intercept
type = sp(sph)(row col);
parms (0 to 10 by 2.5)
(1 to 10 by 3 );
ods select ParmSearch CovParms IterHistory
ConvergenceStatus FitStatistics LRT;
run;
/*-------------------------------------------------------*/
/*-------------------------------------------------------*/
/*---Output 11.2 ---*/
title 'no nugget - exponential covariance model';
proc mixed data=spatvar;
model yield = ;
repeated / subject = intercept
type = sp(exp)(row col);
parms (0 to 10 by 2.5)
(1 to 10 by 3 );
ods select CovParms FitStatistics;
run;
proc mixed data=spatvar;
model yield = ;
repeated / subject = intercept
type = sp(gau)(row col);
parms (0 to 10 by 2.5)
(1 to 10 by 3 );
ods select CovParms FitStatistics;
run;
/*-------------------------------------------------------*/
/*-------------------------------------------------------*/
/*---Output 11.3 ---*/
title 'rcb error model';
proc mixed data=spatvar;
class rep;
model yield = ;
random rep;
ods select CovParms FitStatistics;
run;
title 'incomplete block error model';
proc mixed data=spatvar;
class bloc;
model yield = ;
random bloc;
ods select CovParms FitStatistics;
run;
/*-------------------------------------------------------*/
/*-------------------------------------------------------*/
/*---Outpuyt 11.4 a ---*/
title 'nugget effect - exponential covariance model';
proc mixed data=spatvar;
model yield = ;
repeated / subject = intercept
type = sp(exp)(row col)
local;
parms (0 to 10 by 2.5 )
(1 to 10 by 3 )
(0.05 to 1.05 by 0.25);
ods select IterHistory ConvergenceStatus CovParms;
run;
/*-------------------------------------------------------*/
/*-------------------------------------------------------*/
/*---Outpuyt 11.4 b ---*/
title 'nugget effect - exponential covariance model';
proc mixed data=spatvar convg=1e-7;
model yield = ;
repeated / subject = intercept
type = sp(exp)(row col)
local;
parms (0 to 10 by 2.5 )
(1 to 10 by 3 )
(0.05 to 1.05 by 0.25);
ods select IterHistory ConvergenceStatus CovParms;
run;
/*-------------------------------------------------------*/
/*-------------------------------------------------------*/
/*---Output 11.5 ---*/
data spatvar; set spatvar;
obs = _n_;
run;
proc mixed data=spatvar convg=1e-7;
class obs;
model yield = ;
random obs / type=sp(exp)(row col);
parms (0 to 10 by 2.5 )
(1 to 10 by 3 )
(0.05 to 1.05 by 0.25);
ods select Dimensions IterHistory ConvergenceStatus
FitStatistics CovParms;
run;
/*-------------------------------------------------------*/
--
Muhammad Yaseen
More information about the R-sig-mixed-models
mailing list