All 100 open data a:\MA2PP.DAT data(org=obs) 1 100 date why cal 1 1 5 source C:Bjident.src * * BJIDENT * Box-Jenkins model identification * * Syntax: * @BJIDENT series start end * * Options * DIFF=maximum regular differencings[0] * SDIFFS=maximum seasonal differencings[0] * TRANS=[NONE]/LOG/ROOT * Transformation to apply to data * [GRAPH]/NOGRAPH * Do High-resolution graphs? * SPAN=seasonal span * PROC BJIDENT SERIES START END (01.0016) TYPE SERIES SERIES (01.0016) TYPE INTEGER START END (01.0016) * (01.0016) OPTION INTEGER DIFF 0 (01.0028) OPTION INTEGER SDIFF 0 (01.0040) OPTION CHOICE TRANS 1 NONE LOG ROOT (01.0055) OPTION SWITCH GRAPH 1 (01.0067) OPTION INTEGER SPAN (01.0075) (01.0075) LOCAL INTEGER NBEG NEND SPANL I J (01.0080) LOCAL SERIES CORRS PARTIALS (01.0116) LOCAL SERIES TRANSFRM DIFFED (01.0152) (01.0152) inquire(series=series) nbeg>>start nend>>end (01.0189) inquire(seasonal) spanl>>span (01.0207) (01.0207) if TRANS == 1 (02.0222) set transfrm nbeg nend = series (02.0255) else if TRANS == 2 (02.0272) set transfrm nbeg nend = log(series) (02.0306) else (02.0308) set transfrm nbeg nend = sqrt(series) (02.0342) (02.0342) do i=0,diff (02.0367) do j=0,sdiff (03.0392) diff(diff=i,sdiff=j) transfrm nbeg+i+spanl*j nend diffed (03.0456) disp(store=dispstr) i 'Regular' j 'Seasonal' (03.0501) corr(print=.not.graph,number=20,partial=partials) $ (03.0530) diffed nbeg+i+spanl*j nend corrs (03.0573) if (graph) { (05.0585) graph(max=1.0,min=-1.0,header=dispstr,style=bargraph,$ (05.0615) number=0,key=loright,nodates) 2 (05.0632) # corrs 1 21 (05.0652) # partials 1 21 (05.0672) } (04.0672) enddo j (02.0674) enddo i (01.0676) end @Bjident why corr(par=pwhy) why / res Correlations of Series WHY 5/Year Data From 1:01 To 20:05 Autocorrelations 1: -0.2159559 -0.1394682 -0.0333076 -0.0906771 -0.0243312 -0.0080972 7: -0.0786371 0.0634329 0.1142907 -0.0390795 0.0372472 -0.1048656 13: -0.0276452 0.1619092 -0.0050642 -0.0377746 -0.1194449 0.0015802 19: 0.0266008 -0.1433733 0.2464700 -0.0785214 0.0379128 -0.0109844 25: 0.0775782 Partial Autocorrelations 1: -0.2159559 -0.1952091 -0.1215470 -0.1747833 -0.1395705 -0.1277422 7: -0.2033381 -0.1020918 0.0132485 -0.0545674 0.0131309 -0.1148677 13: -0.0887987 0.1079795 0.0776707 0.0521760 -0.1159840 -0.0618553 19: -0.0453421 -0.2382501 0.1613602 -0.0976170 -0.0182834 -0.0924417 25: 0.0858608 graph 1 #why Box(AR=2,constant) why / res Dependent Variable WHY - Estimation by Box-Jenkins Iterations Taken 4 5/Year Data From 1:03 To 20:05 Usable Observations 98 Degrees of Freedom 95 Centered R**2 0.086549 R Bar **2 0.067318 Uncentered R**2 0.997510 T x R**2 97.756 Mean of Dependent Variable 40.002481633 Std Error of Dependent Variable 2.102228683 Standard Error of Estimate 2.030236966 Sum of Squared Residuals 391.57690297 Durbin-Watson Statistic 2.009606 Q(24-2) 28.537435 Significance Level of Q 0.15854692 Variable Coeff Std Error T-Stat Signif ******************************************************************************* 1. CONSTANT 39.99144024 0.13976332 286.13687 0.00000000 2. AR{1} -0.26531477 0.10202945 -2.60037 0.01079931 3. AR{2} -0.20229496 0.10310116 -1.96210 0.05267696 Dependent Variable WHY - Estimation by Box-Jenkins Iterations Taken 3 5/Year Data From 1:02 To 20:05 Usable Observations 99 Degrees of Freedom 97 Centered R**2 0.049472 R Bar **2 0.039672 Uncentered R**2 0.997435 T x R**2 98.746 Mean of Dependent Variable 40.005537374 Std Error of Dependent Variable 2.091696510 Standard Error of Estimate 2.049785438 Sum of Squared Residuals 407.55717333 Durbin-Watson Statistic 2.063728 Q(24-1) 27.482845 Significance Level of Q 0.23597150 Variable Coeff Std Error T-Stat Signif ******************************************************************************* 1. CONSTANT 39.99237012 0.16822523 237.73110 0.00000000 2. AR{1} -0.22511063 0.10018783 -2.24689 0.02691413 Dependent Variable WHY - Estimation by Box-Jenkins Iterations Taken 13 5/Year Data From 1:01 To 20:05 Usable Observations 100 Degrees of Freedom 97 Centered R**2 0.171171 R Bar **2 0.154082 Uncentered R**2 0.997740 T x R**2 99.774 Mean of Dependent Variable 39.976883000 Std Error of Dependent Variable 2.100739803 Standard Error of Estimate 1.932130208 Sum of Squared Residuals 362.11333261 Durbin-Watson Statistic 1.861594 Q(25-2) 24.234943 Significance Level of Q 0.39089569 Variable Coeff Std Error T-Stat Signif ******************************************************************************* 1. CONSTANT 39.97966368 0.04439579 900.52828 0.00000000 2. MA{1} -0.43385250 0.09810747 -4.42222 0.00002553 3. MA{2} -0.34717254 0.09850062 -3.52457 0.00064909 graph 2 #why #res @bjident res corr(par=par) res Correlations of Series RES 5/Year Data From 1:01 To 20:05 Autocorrelations 1: -0.0223217 -0.0611014 -0.1689748 -0.1660033 -0.0821312 -0.0407867 7: -0.0590099 0.0739497 0.1339995 -0.0060395 0.0098337 -0.1031251 13: -0.0015111 0.1606093 0.0063829 -0.0612192 -0.1562871 -0.0699525 19: 0.0165771 -0.0943197 0.2330209 -0.0316931 0.0855895 -0.0314223 25: 0.0600032 Partial Autocorrelations 1: -0.0223217 -0.0616304 -0.1725394 -0.1859829 -0.1301793 -0.1219203 7: -0.1717799 -0.0459743 0.0432989 -0.0782532 -0.0305725 -0.1064013 13: -0.0147813 0.1594648 0.0246791 -0.0417257 -0.1465036 -0.0821903 19: -0.0322341 -0.2006624 0.1770934 -0.1384515 -0.0302379 -0.0735752 25: 0.0959770 Box(AR=2,constant) why / res2 Dependent Variable WHY - Estimation by Box-Jenkins Iterations Taken 4 5/Year Data From 1:03 To 20:05 Usable Observations 98 Degrees of Freedom 95 Centered R**2 0.086549 R Bar **2 0.067318 Uncentered R**2 0.997510 T x R**2 97.756 Mean of Dependent Variable 40.002481633 Std Error of Dependent Variable 2.102228683 Standard Error of Estimate 2.030236966 Sum of Squared Residuals 391.57690297 Durbin-Watson Statistic 2.009606 Q(24-2) 28.537435 Significance Level of Q 0.15854692 Variable Coeff Std Error T-Stat Signif ******************************************************************************* 1. CONSTANT 39.99144024 0.13976332 286.13687 0.00000000 2. AR{1} -0.26531477 0.10202945 -2.60037 0.01079931 3. AR{2} -0.20229496 0.10310116 -1.96210 0.05267696