Predictive Analytics and Business Forecasting- Time series regression

FILENAME REFFILE '/home/u49607241/Assignment_Tejdeep/GLB.Ts+dSST.csv'; /* PROC IMPORT DATAFILE=REFFILE DBMS=CSV OUT=WORK.IMPORT3; GETNAMES=YES; RUN; PROC CONTENTS DATA=WORK.IMPORT; RUN; */ /* Durbin Watson Test*/ proc reg data=WORK.IMPORT3; model CO2= / dw; run; quit; /* Linear Regression Y=mX+C */ proc reg data=WORK.IMPORT3 alpha=0.05 plots(only)=(diagnostics residuals fitplot observedbypredicted); model CO2=Date /; run; quit; /* Linear Regression Y=m1 X+m2 X^2 + C */ proc glmselect data=WORK.IMPORT; model CO2=Date Date*Date / showpvalues selection=none; run; proc reg data=WORK.IMPORT3 alpha=0.05 plots(only)=(diagnostics residuals observedbypredicted); ods select DiagnosticsPanel ResidualPlot ObservedByPredicted; model CO2=&_GLSMOD /; run; quit; /* ARIMA Exploratory Analysis to find p,d,q */ proc sort data=WORK.IMPORT3; by Date; run; proc timeseries data=WORK.IMPORT3 seasonality=12 plots=(series histogram cycles corr); id Date interval=month; var CO2 / accumulate=none transform=none dif=0 sdif=0; run; /* ARIMA Prediction using defined p,d,q */ ods noproctitle; ods graphics / imagemap=on; proc arima data=WORK.IMPORT3 plots (only)=(series(corr crosscorr) residual(corr normal) forecast(forecast forecastonly)); identify var=CO2(1); estimate p=(1 2) q=(1) method=ML; forecast lead=12 back=0 alpha=0.05 id=Date interval=month; outlier; run; quit;