Data coal_1;
input price@@;
difprice=dif(price);
time=intnx('month','1jan2003'd,_n_-1);
format time monyy.;
cards;
272 274 273 272 272 269 267 267 269 269 277 279
290 290 290 297 326 340 340 332 330 333 429 438
450 455 451 450 450 450 450 437 433 449 445 445
450 459 460 456 454 450 446 455 460 463 474 493
511 503 493 480 475 485 490 495 510 525 531 539
596 656 645 636 685 860 1000 940 940 938 808 624
627 598 585 595 607 602 590 595 614 642 725 780
810 815 733 729
;
proc gplot;
plot difprice*time;
symbol v=star c=red i=join;
proc arima;
identify var=price(1) minic p=(0:5) q=(0:5);
estimate q=1 noint;
forecast lead=10 id=time out=results;
proc gplot data=results;
plot price*time=1 forecast*time=2 l95*time=3 u95*time=3/overlay;
symbol1 c=blue i=none v=star;
symbol2 c=red i=join v=none l=1 w=1;
symbol3 c=green i=join v=none l=2 w=2;
run;
输出结果:
Forecasts for variable price
Obs Forecast Std Error 95% Confidence Limits
89 750.4078 36.0938 679.6652 821.1503
90 750.4078 66.5547 619.9629 880.8526
91 750.4078 86.9270 580.0340 920.7815
92 750.4078 103.3589 547.8280 952.9875
93 750.4078 117.5152 520.0821 980.7334
是程序错误,还是模型出问题了,不是凑巧吧,上下置信区间去不一样?%>_<%