How to Show MAPE(Mean Absolute Percent error) in the Proc Arima model by default

critiasun — Thu, 28 May 2020 01:20:28 GMT

data arima.js_dispos;
set arima.js_2005to2018;
label year="y" cases="c";
run;

proc gplot data=arima.js_dispos;

plot cases*year;
symbol i=spline v=star ci=red cv=bib;
run;

proc arima data=arima.js_dispos;
identify var=cases stationarity=(adf=1);
run;

data arima.js_dispos_dif2;
set arima.js_dispos;
z=cases;
difz=dif(z);
difz2=dif(difz);
label z="n" difz="d1" difz2="d2" ;
run;

proc gplot data=arima.js_dispos_dif2;
plot (z difz difz2)*year;
symbol i=spline v=star ci=red cv=bib;
run;
proc arima data=arima.js_dispos_dif2;
identify var=difz2 stationarity=(adf=1);
run;

proc arima data=arima.js_dispos_dif2;
identify var=difz2 stationarity=(adf=1) WHITENOISE=IGNOREMISS;
run;

proc arima data=arima.js_dispos_dif2;
identify var=difz2 minic perror=(8:11);
run;

/*test proc arima data=arima.js_dispos_dif2;
identify var=difz2 minic P=(0:5) Q=(0:5) ;
run;

---------------test end*/

proc arima data=arima.js_dispos_dif2;
identify var=z(2) nlag=6;
estimate q=0 p=0 method=ml;
forecast lead=5 out=arima.outvalue id=year;
run;
quit;

topic How to Show MAPE(Mean Absolute Percent error) in the Proc Arima model by default in SAS Forecasting and Econometrics

How to Show MAPE(Mean Absolute Percent error) in the Proc Arima model by default