I think I found what was wrong : if I do scaling on the matices y and xreg,
all seems good. Here is what I do :
> dd<-read.csv2("c:/modelfr2.csv")
> y<-dd[,1]
> x<-dd[,-1:-2]
> x1<-scale(x)
> y1<-scale(as.matrix(as.double(y)))
> ari<-arima0(y1,xreg=as.matrix(x1),order=c(1,0,3))
and I got :
> ari$coef
ar1 ma1 ma2 ma3
intercept
0.981959774 -0.133513133 -0.313765424 -0.120906582 -0.014087163
TempLT17 lag2 We hivergaz
lag7
0.595713158 0.081242497 -0.074302282 0.121198338
0.049151266
octobre vendredi Tmoycarre Tmaxi
avril
-0.046997365 -0.024616056
0.401650456 -0.067899417 -0.062226297
TempGT17 vacances Lineartrend TempLT171
feriesemaine
-0.174204375 -0.010849255 -1.593994601
0.094049181 -0.015111721
vacoct MoistransT17 jeudi mercredi
Tmoycarre1
0.002894903 0.066673677 -0.008290431 -0.004808032
0.048113321
WET17 vacdec MoistransTnx septembre
MoistransDeltaT
-0.018610689 -0.021541452 0.028866118 -0.060574403
0.024856178
MoistransDTmax Rechauffement samedi janvier
automne
-0.011687046 0.004914881 -0.007455964 -0.008352493
0.037430983
periode quadratrend aout tempeff
MoistransDTmin
-0.812194507
1.614215908 -0.004387879 -0.050769121 -0.005031441
MoistransTmin hiver juin
0.026752595 0.027127835 0.007552805
> sqrt(diag(ari$var.coef))
ar1 ma1 ma2 ma3
intercept
0.010172128 0.042899460 0.051893431 0.039494948
0.057324206
TempLT17 lag2 We hivergaz
lag7
0.020040066 0.027233295 0.004830991 0.028450063
0.015719575
octobre vendredi Tmoycarre Tmaxi
avril
0.023873606 0.003941548 0.033016873 0.013613845
0.019573895
TempGT17 vacances Lineartrend TempLT171
feriesemaine
0.016457262 0.008907412 0.401197993 0.020490511
0.002024213
vacoct MoistransT17 jeudi mercredi
Tmoycarre1
0.005764023 0.015146301 0.003811101 0.003136478
0.021529475
WET17 vacdec MoistransTnx septembre
MoistransDeltaT
0.004133053 0.007681165 0.021842469 0.023770587
0.005238365
MoistransDTmax Rechauffement samedi janvier
automne
0.003566694 0.001953077 0.002418192 0.012942518
0.015704967
periode quadratrend aout tempeff
MoistransDTmin
0.208897607 0.413539315 0.011196934 0.043177465
0.003314960
MoistransTmin hiver juin
0.026484594 0.016473332 0.008654350
The only problem now is : how can I retrieve now the true coefficients and
the true se for the non scaled regressors and intercept ?
Can anyone help me ?
Thank you very much.
Pascal Grandeau
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._