/* MARK 9650 Project 2: Logistic Regression on Lasagne data          */
/* Dependent Variable: HAVETRIED									 */

OPTIONS PS=52;
OPTIONS LS=78;
OPTIONS NODATE;
OPTIONS PAGENO=1;

DATA ANALYSIS;
   INFILE 'A:/LAS_ANALY.PRN';
   INPUT TRAN_AGE TRAN_WEIGHT TRAN_INC GENDER LIVEALONE HAVETRIED;
RUN;

PROC LOGISTIC DATA=ANALYSIS DESCENDING;
   MODEL HAVETRIED=TRAN_AGE TRAN_WEIGHT TRAN_INC GENDER LIVEALONE
												/ SELECTION=STEPWISE;
   OUTPUT OUT=OUTHAT P=PRED;
RUN;

DATA NEWANALY;
   SET OUTHAT;
   IF PRED > .56 THEN CLASSIF_REG=1;
   IF PRED < .56 THEN CLASSIF_REG=0;
   IF PRED > .675 THEN CLASSIF_BIAS=1;
   IF PRED < .675 THEN CLASSIF_BIAS=0;
RUN;

PROC FREQ DATA=NEWANALY;
   TITLE "Classification Table with PRED cutoff of .56 (Analysis Sample)";
   TABLE HAVETRIED*CLASSIF_REG;
RUN;

PROC FREQ DATA=NEWANALY;
   TITLE "Classification Table with 15% misclassified non-triers (Analysis Sample)";
   TABLE HAVETRIED*CLASSIF_BIAS;
RUN;

DATA HOLDOUT;
   INFILE 'A:/LAS_HOLDOUT.PRN';
   INPUT TRAN_AGE TRAN_WEIGHT TRAN_INC GENDER LIVEALONE HAVETRIED;
   Y = exp(0.2103 - 0.8898*TRAN_AGE + 0.9434*TRAN_INC + 1.1608*LIVEALONE);
   PRED = Y/(1+Y);
   IF PRED > .56 THEN CLASSIF_REG=1;
   IF PRED < .56 THEN CLASSIF_REG=0;
   IF PRED > .675 THEN CLASSIF_BIAS=1;
   IF PRED < .675 THEN CLASSIF_BIAS=0;
RUN;

PROC FREQ DATA=HOLDOUT;
   TITLE "Classification Table with PRED cutoff of .56 (Holdout Sample)";
   TABLE HAVETRIED*CLASSIF_REG;
RUN;

PROC FREQ DATA=HOLDOUT;
   TITLE "Classification Table with 15% misclassified non-triers (Holdout Sample)";
   TABLE HAVETRIED*CLASSIF_BIAS;
RUN;