Hi all,
I have I got it. I see the way on my mind. I will try this week end and if
I get it I will send.
Thanks.
2012/5/10 Trying To learn again <tryingtolearnag...@gmail.com>
> Hi all,
>
> I want to make the following:
>
> I want to run a linear regression on each column of a matrix "estima" on
> the correspondent column on the matrix "estima2".
>
> You see I want to regress estima[,1] on estima2[,1] this way to all
> columns....
>
> At the same time I want to make a regression adding each time a new
> observation.
>
> You see, the first regression will regress only one observation with one
> observation (I now this has no sense in this only one observation step)
>
> the second turn of observation will make
>
> estima[1:2,n] on estima2[1:2,n] for all "n".
>
> Third stimation will make
>
> estima[1:3,n] on estima2[1:3,n] for all "n".
>
> And so on.
>
> Make this, I want to make an output matrix on each "t-value" associated
> with the "regressor".
>
> Conclusion my final matrix called "t value" should include al the t values
> on the regression each of them incorporating a new observation, with the
> same rows and colums than "estima".
>
> I have tried several thing but I cannot achive.
>
> I writte to see if you can guide me¡¡¡
>
> I swear I´m trying.
>
>
>
> randz<-matrix(rnorm(5000),50,100)
>
> H<-matrix(0,50,100)
>
> H[1,]<-randz[1,]
> for (i in 2:50){
> if(i < 26) {
> H[i, ] <- 0.6 * H[i-1, ] + randz[i, ]
> } else {
> H[i, ] <- H[i-1, ] + randz[i, ]
> }
> }
>
>
> write.table(H, file = "datad.txt")
> g<-read.table("datad.txt")
>
> hy<-nrow(g)-1
> estima<-H[2:nrow(g), ]
> estima2<-H[ 1:hy, ]
>
> mycoef <- function (x,y)
> a<-estima
> b<-estima2
> f<-summary(lm(a~b))
> ff<-coef(f)
> ff[2,"t value"]
> tvalue <- sapply (2:ncol(b) , function (i){
> y<-a[,i]
> x<-b[,i]
> mycoef(x,y)
>
> }
> )
> print (summary(tvalue))
>
>
>
[[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.
