数据标准化点滴

记数据列为变量，行为观测

1 依据观单独的测值均值作标准化，公式形如：（变量值-观测均值）/观测均值

data stocks;
   set sashelp.stocks;
array m{*}  open high low close adjclose;
array m_mean{*} open high low close adjclose;
   m_avg= mean (of open high low close adjclose);
do i=1 to dim(m);
   m_mean{i}=((m{i}-m_avg)/m_avg);
end;
drop i;
run;
2 根据某个条件下的所有观测值均值作标准化，该条件可以为时期、分类变量等等

data stocks1;
   set sashelp.stocks;
run;
proc means data=stocks1 ;
   var open high low close adjclose;
   class stock;
   output out=m_mean mean= / autoname;
run;
data stocks1;
   set stocks1 ;
   if stock="IBM" THEN
   open=open/102.65;
run;

 3 PROC STDIZE Procedure

PROC STDIZE < options > ;
   BY variables ;
   FREQ variable ;
   LOCATION variables ;
   SCALE variables ;
   VAR variables ;
   WEIGHT variable ;