options ps=54 ls=76;

data q4_2002;
infile 'A:\q4_2002.out';
input   city1    $
        city2    $
        ave_fare
        distance
        passengers;
x=ranuni(1234);

run;

proc sort;
by x;
run;

data q4_2002a;
set q4_2002;
case=_n_;
if case<=500 then do;
yval1=ave_fare;
yval2=.;
end;
if case> 500 then do;
yval1=.;
yval2=ave_fare;
end;
run;

proc reg data=q4_2002a;
model yval1=distance;
output out=val1 p=yhatv1;
run;

data val1; set val1;
eval1=ave_fare-yhatv1;
run;

proc reg data=q4_2002a;
model yval2=distance;
output out=val2 p=yhatv2;
run;

data val2; set val2;
eval2=ave_fare-yhatv2;
run;

proc means mean std uss data=val1;
where case > 500;
var eval1;
run;

proc means mean std uss data=val2;
where case <= 500;
var eval2;
run;

quit;