Ʈ  |  Contact Us
 
Ȩ > SAS Tech & Tip > SAS Programming
[SAS α׷ ǽ] ƽ ȸ͸ ̿ PROC PLM ǽ 2019.05.22
ڼ 525 1
http://www.mysas.co.kr/sas_tiptech/a_question.asp?b_no=11108&cmd=content&bd_no=5

: SAS PROC PLM

ƽ ȸͺм 䱸˴ϴ.

 

 

1. Introduction

ð ȸ ǽغҽϴ. ȸ ϸ鼭, ϰ Ķ͸ ϰ, ڵ带 ϴ Ĩϴ. м 忡 ϱ ---- μ Ұϰ, մϴ. ǻ з ϱ ؾ . PLM ̷ ȿ ݴϴ. , ĸ𵨸 մϴ.

 

2. Assessment

PROC PLM ũ 4 ֽϴ.

  1. SCORE Ͽ ο Ϳ Scoring

  2. EFFECTPLOT ðȭ

  3. ESTIMATE, LSMEANS, SLICE TEST Parameter

  4. SHOW

PROC PLM ϱ ؼ м Procedure STORE ؾ մϴ. STORE ϴ Ͻ [GEE, GENMOD, GLIMMIX, GLM, GLMSELECT, LIFEREG, LOGISTIC, MIXED, ORTHOREG, PHREG, PROBIT, SURVEYLOGISTC, SURVEYPHREG ] ֽϴ. ?

 

3. USAGE

, PROC PLM غڽϴ. STORE, EFFECTPLOT, SHOW ǽغڽϴ.

3.1 STORE

<code1>

Data Neuralgia;

   input Treatment $ Sex $ Age Duration Pain $ @@;

   datalines;

P F 68  1 No  B M 74 16 No  P F 67 30 No  P M 66 26 Yes B F 67 28 No  B F 77 16 No A F 71 12 No  B F 72 50 No  B F 76  9 Yes A M 71 17 Yes A F 63 27 No  A F 69 18 Yes B F 66 12 No  A M 62 42 No  P F 64  1 Yes A F 64 17 No  P M 74  4 No  A F 72 25 No P M 70  1 Yes B M 66 19 No  B M 59 29 No  A F 64 30 No  A M 70 28 No  A M 69  1 No B F 78  1 No  P M 83  1 Yes B F 69 42 No  B M 75 30 Yes P M 77 29 Yes P F 79 20 Yes A M 70 12 No  A F 69 12 No  B F 65 14 No  B M 70  1 No  B M 67 23 No  A M 76 25 Yes P M 78 12 Yes B M 77  1 Yes B F 69 24 No  P M 66  4 Yes P F 65 29 No  P M 60 26 Yes A M 78 15 Yes B M 75 21 Yes A F 67 11 No  P F 72 27 No  P F 70 13 Yes A M 75  6 Yes B F 65  7 No P F 68 27 Yes P M 68 11 Yes P M 67 17 Yes B M 70 22 No  A M 65 15 No P F 67  1 Yes A M 67 10 No  P F 72 11 Yes A F 74  1 No  B M 80 21 Yes A F 69  3 No

;

run;

ȯ Ű ͷ ǽ ϰڽϴ. Treatment ġ (Drug A, Drug B, Placebo), , , ӵ Ⱓ, Դϴ. ͷ ϰ Logistic ȸ͸ ô.

<code 2>

title 'Logistic Model on Neuralgia';

proc logistic data=Neuralgia;

   class Sex Treatment;

   model Pain(Event='Yes')= Sex Age Duration Treatment;

/* ְڽϴ */

   store PainModel / label='Neuralgia Study';

/*̷ м ٸ store ֽ߰ø ˴ϴ. */

run;

PROC Logistic Ͽ ̷ پ 跮 Ȯ ֽϴ. ִ쵵 , , Odds Ratio  Ǵ .. ⿡ ڼ ٷ ʰڽϴ.

ڵ带 , ƽ α׷ ߴٰ մϴ. ׷ ο ȯ Ͱ ߰Ǿϴ. 𵨷 ȯڰ ˾Ƴ ͽϴ. ̶ PROC PLM մϴ.

 

 

<code3>

proc plm restore=PainModel; /*restore ȣմϴ */

          score data=NewPatients

          out=NewScore predicted

          LCLM UCLM / ilink; /* ILINK ɼ */

run;

 

ILINK ɼǿ Ǵ computes and displays estimates and standard errors of LS-means (not differences) on the inverse linked scale ɴϴ. ּ ǥؿ ִ ⺻ ɼԴϴ.

 

PLM ϳԴϴ.  ȣƴ, Լ  ߰ ð .. ͸ ʾƵ ɴϴ.

 

proc print data=NewScore;

run;


 

PLM NewScore ? ο Ϳ м ֽϴ.

LCLM UCLM ɼ ߰ؼ ׷ ŷѰ ŷڻѵ µǾϴ. Ilick ɼ п Logistic Ȯ Խϴ.

 

1 ȯ 1% Ȯ ̰, 3 ȯ 63.8% Ȯ ٴ Խϴ. 3 ȯڴ ؾ߰ڳ׿.

 

3.2 EFFECTPLOT

EFFECTPLOT LOGISTIC GENMOD Procedure մϴ. EFFECTPLOT Ͽ ġ Ȯ ðȭ غڽϴ.

 

/* 2. Effect plot */

proc plm restore=PainModel;

effectplot

slicefit(x=Age sliceby=Treatment plotby=Sex);

/* slicefit F test մϴ*/

 

run;

 

 

µ ׷ ؼغ, ̰ Ȯ ׿. Placebo  ̿ , A B ȿ ƺԴϴ. ̿ Ȯ ϱ.

 

3.3 Estimate, SHOW

/* 3. Exp estimate ̸ мմϴ.  */

proc plm restore=PainModel;

   /* 'Exponentiated' ÷ ġ Ÿϴ  */

   estimate 'Pairwise A vs B' Treatment 1 -1 / exp CL;

run;

 

 

/* 4. ⺻ 跮 л */

proc plm restore=PainModel;

   show Parameters COVB Program;

run;

 

SHOW ɼǿ ⺻ α׷ ȣǾ׿. ȣ м α׷ ־ ϴ. COVB ؼ л ҷ ֽϴ.

 

 

 

4. Summary

ó PROC PLM پ ȸ м ϰ ȣϿ ִٴ 𵨸 ϰ ֽϴ. ϰ, ðȭϰ, ϴ پ м ȿ Ͻñ ٶ ̹ ġڽϴ.

 

 

Ps.

Ͻô ̳ ۷ ޾ ֽø ϰڽϴ.