Hi, everyone.
I have some questions about quantile regression in R.
I am running an additive quantile regression first for a complete matrix and
then with some selected rows.
I am doing the following:
datos <-read.table("Regresion multiple.txt",header=T)
Fit<-rqss(datos$campings
~datos$Cobarbogrupo+datos$CobSDgrupo+datos$Areadecultivosgrupo, tau=0.9)
summary.rq(Fit)
#The output is:
Coefficients:
Value Std. Error
t value Pr(>|t|)
(Intercept) 1.46561 0.35857
4.08733 0.00004
datos$Cobarbogrupo 0.09357 0.05064
1.84784 0.06473
datos$CobSDgrupo -0.07328 0.01671
-4.38629 0.00001
datos$Areadecultivosgrupo 0.02706 0.02202 1.22873
0.21927
Warning message:
In summary.rq(Fit) : 1 non-positive fis
#So far, so good. Now I select a group of rows I am interested in as follows:
sub<-subset(datos, Regtur==1)
write.table(sub, "Regtur1.txt")
#And run the quantile regression again:
dato<- read.table("Regtur1.txt", header=T)
Fit<-rqss(dato$campings ~
dato$Cobarbogrupo+dato$CobSDgrupo+dato$Areadecultivosgrupo, tau=0.9)
summary.rq(Fit)
#As you can see, the commands are the same. But, I get this output:
Coefficients:
coefficients lower bd
upper bd
(Intercept) 2.67500 -2.58710
8.81632
dato$Cobarbogrupo 1.00000 0.66909
3.01845
dato$CobSDgrupo 0.32500 -0.56172
1.77540
dato$Areadecultivosgrupo -0.15000 -0.43830
0.12112
Warning message:
In rq.fit.br(x, y, tau = tau, ci = TRUE, ...) : Solution may be nonunique
My questions are these:
1) Why do I get different outputs? Why I don't get the p values for each
coefficient in the second output? Is there a problem with tables?
2) Are the two outputs showing the same information? Can I interpret that if
zero is included in the interval of the second output, the coefficient is not
different from zero?
Thanks in advance to anyone who can help me.
Federico
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
