data newlib.abc;
input a1 @@;
day=intnx( 'day','01mar2007'd,_n_-1);
format day yymmdd10.;
cards;
102.7 181.7 79 126.01 55.3 142.2 63.2 134.3
47.4 126.4 86.9 23.7 39.5 94.8 63.2 31.6
71.1 79 39.5 63.2 55.3 7.9 79 142.2
71.1 63.2 102.7 126.4 126.4 158 165.9 110.6
102.7 150.1 94.8 102.7 134.3 71.1 110.6 134.3
126.4 63.2 94.8 221.2 63.2 118.5 102.7 134.3
94.8 71.1 142.2 94.8 181.7 134.3 86.9 173.8
110.6 221.2 126.4 173.8 165.9 86.9 134.3 118.5
94.8 158 86.9 63.2 71.1 79 158 158
55.3 39.5 118.5 63.2 63.2 118.5 110.6 126.4
39.5 63.2 110.6 102.7 165.9 134.3 63.2 110.6
63.2 102.7 102.7 150.1 94.8 94.8 94.8 94.8
71.1 110.6 158 118.5 79 110.6 63.2 94.8
63.2 110.6 110.6 86.9 94.8 63.2 55.3 102.7
;
run;

proc gplot data=newlib.abc;
symbol i=spline v=star h=2 c=red;
plot a1*day;
run;

proc arima data=newlib.abc;
identify var=a1 nlag=24;
run;

estimate p=2 q=2;
run;

forecast lead=6 interval=day id=day out=out;
run;
proc print data=out;
run;

怎么这个预测出来的效果不好啊,相对误差太大拉

谁能帮帮我啊,很急啊,谢谢!!!!!!!