Saludos
Estoy utilizando el PROC GLIMMIX para ajustar un modelo conjunto
data WORK.IMPORT_V;
length dist $7;
set WORK.IMPORT;
response=(VAR16=1);
dist="Binary";
output;
response = Rep;
dist="Poisson";
output;
keep APR APT Diagnostico Cursos IDs Escuela response dist ;
run;
proc glimmix data= work.import_v (where= (dist="Binary"));
class IDs Escuela Cursos;
model response(event='1') = APR APT Diagnostico / s dist=binary;
run;
proc glimmix data=work.import_v(where=(dist="Poisson"));
class IDs Escuela Cursos;
model response = APR APT Diagnostico / s dist=Poisson;
run;
proc glimmix data=work.import_v;
class dist;
class IDs Escuela Cursos;
model response(event='1') = dist dist*APR dist*APT dist*Diagnostico /
noint s dist=byobs(dist);
run;
proc glimmix data=work.import_v;
class IDs dist;
class Escuela Cursos;
model response(event='1') = dist dist*APR dist*APT dist*Diagnostico /
noint s dist=byobs(dist);
random int / subject=IDs;
run;
El modelo conjunto consta de dos submodelos, un logístico y otra poisson. La salida de dicho modelo me dice que la matriz G estima que no es claramente positiva. Adjunto imagen con la salida.
La verdad no se como interpretar esta salida. Si tienes la oportunidad, también me gustaría contar con apoyo literario ya que no encuentro suficiente informacion del tema para aclarar mis dudas.
Mis datos se encuentran organizados de la siguiente manera:
Identificaciones | Cursos | Notas | Reps | Sexo | Ano | Escuela | Distrito | IGS | APTO | ABR | Nivel1 | Nivel2 | Nacimiento | Diagnostico | VAR16 |
00BA1F4E0C | MATE_3171 | W | 1 | F | 2013 | Privada | Mayaguez | 323 | 544 | 542 | 0 | 0 | 1995 | 0 | 0 |
00D72CEB8A | MATE_3171 | F | 2 | F | 2013 | Publica | Mayaguez | 305 | 531 | 528 | 0 | 0 | 1995 | 0 | 0 |
00A362D565 | MATE_3171 | W | 1 | F | 2013 | Publica | Mayaguez | 305 | 531 | 528 | 0 | 0 | 1995 | 0 | 0 |
00AA00E246 | MATE_3172 | hacer | 1 | F | 2013 | Publica | Mayaguez | 305 | 531 | 528 | 0 | 0 | 1995 | 0 | 1 |
00CD4D51FA | MATE_3171 | hacer | 1 | METRO | 2013 | Publica | Mayaguez | 327 | 589 | 584 | 0 | 2 | 1994 | 0 | 1 |
00CD4D51FA | MATE_3172 | segundo | 2 | METRO | 2013 | Publica | Mayaguez | 327 | 589 | 584 | 0 | 2 | 1994 | 0 | 1 |
00CD4D51FA | MATE_3031 | W | 1 | METRO | 2013 | Publica | Mayaguez | 327 | 589 | 584 | 0 | 2 | 1994 | 0 | 0 |
Agradezco su amable atención y colaboración.
I have never used the DIST=BYOBS option, but sometimes a singular G matrix can mean that the data do not fit the model. I see that the
dist*APT and dist*Diagnostico effects
are not significant. Try removing those effects from the model.
Calling all data scientists and open-source enthusiasts! Want to solve real problems that impact your company or the world? Register to hack by August 31st!
ANOVA, or Analysis Of Variance, is used to compare the averages or means of two or more populations to better understand how they differ. Watch this tutorial for more.
Find more tutorials on the SAS Users YouTube channel.