/* MARK 9650 Project 2b: Discriminant Analysis on Lasagne data        			*/
/* Dependent Variable: Have Tried with non-triers misclassified 15% of the time */

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 SORT;
BY HAVETRIED;
RUN;

PROC MEANS;
CLASS HAVETRIED;
VAR TRAN_AGE TRAN_WEIGHT TRAN_INC GENDER LIVEALONE;
RUN;

PROC GLM;
CLASS HAVETRIED;
MODEL TRAN_AGE TRAN_WEIGHT TRAN_INC=HAVETRIED;
RUN;

PROC FREQ;
TABLES HAVETRIED*GENDER HAVETRIED*LIVEALONE / CHISQ;
RUN;

PROC DISCRIM METHOD=NORMAL POOL=YES MANOVA LIST;
CLASS HAVETRIED;
PRIORS "0"=.617 "1"=.383;
VAR TRAN_AGE TRAN_INC LIVEALONE GENDER;
RUN;

DATA HOLDOUT;
   INFILE 'A:/LAS_HOLDOUT.PRN';
   INPUT TRAN_AGE TRAN_WEIGHT TRAN_INC GENDER LIVEALONE HAVETRIED;
   Z0 = -1.15178 + 0.48292*TRAN_AGE - 0.43831*TRAN_INC + 0.67312*LIVEALONE
													+ 1.87959*GENDER;
   Z1 = -1.86788 - 0.40802*TRAN_AGE + 0.46493*TRAN_INC + 1.72320*LIVEALONE
													+ 2.22864*GENDER;
   IF Z0 > Z1 THEN Z = 0;
   IF Z1 > Z0 THEN Z = 1;
RUN;

PROC FREQ;
   TABLE HAVETRIED*Z;
RUN;