498
Le logiciel R - Effectuer des analyses statistiques
Pour l’intervalle de confiance de la valeur moyenne du poids des b´ eb´ es pour
un poids de la m` ere de 56 kg :
> predict(modele1,data.frame(LWT=lwt0),interval="confidence")
fit
lwr
upr
1 2916.504 2811.225 3021.783
Repr´ esentons maintenant l’intervalle de confiance et l’intervalle de pr´ evision
pour une s´ erie de nouvelles valeurs de poids de la m` ere :
> x <- seq(min(LWT),max(BWT),length=50)
> intpred <- predict(modele1,data.frame(LWT=x),interval=
+
"prediction")[,c("lwr","upr")]
> intconf <- predict(modele1,data.frame(LWT=x),interval=
+
"confidence")[,c("lwr","upr")]
> plot(BWT∼LWT,xlab="Poids de la mère",ylab="Poids de l’enfant")
> abline(modele1)
> matlines(x,cbind(intconf,intpred),lty=c(2,2,3,3),
+
col=c("red","red","blue","blue"),lwd=c(2,2,1,1))
> legend("bottomright",lty=c(2,3),lwd=c(2,1),
+
c("confiance","prévision"),col=c("red","blue"))
●
●
● ●
●
●
●
●
●
●
●
●
●
●
● ●
●
● ● ● ●
●
●
● ●
●
●
●
●
● ●
●
●
●
●
●
●
●
●
●
●●
●
●
● ●
● ●
●
●
● ● ●
●
●
● ●
●
●
● ● ●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
●
● ● ●
●
●
● ●
●
● ●
●
●
●
● ● ●
●
●
●
●
●
●
● ●
●
●
● ●
●
●
●
● ●
●
● ●
●
●
●
●
●
●
●
●
●
●
●
●
●●
●
●
●
●
●
●
●
●
●
●
●
●
● ● ●
●
●
● ● ●
●
●
●●
●
● ●
●
● ● ●
●
●
● ● ●
●
●
●
● ●
●
●
●
●
● ●
●
●
●
40
60
80
100
1000
2000
3000
4000
5000
Poids de la mère
Poids de l'enfant
confiance
prévision
Figure 12.3 – Visualisation de l’intervalle de confiance et de l’intervalle de
pr´ evision.