Thank you David.
The plot() you sugested worked nicely.
You and Weyland both think ts class objects are difficult to use. I had not
considered doing differently, but I have just gotten many tutorials on package
zoo and how to use it in econometrics. Looks promising, will read everything
now.
But let me ask first, I am going to do some exercises using vars package, and I
wanted to replicate on my own database the following script (below) that I have
managed to put together using the data Canada (available in R):
So I compared the structure of "Canada" and I concluded it was a "ts" object,
and that's why I decided to create my own data the same way. Is this necessary
or am I taking an irrelevant precaution?
> str(Canada)
mts [1:84, 1:4] 930 930 930 931 933 ...
- attr(*, "dimnames")=List of 2
..$ : NULL
..$ : chr [1:4] "e" "prod" "rw" "U"
- attr(*, "tsp")= num [1:3] 1980 2001 4
- attr(*, "class")= chr [1:2] "mts" "ts"
> str(dados2) #my database on public dept
mts [1:210, 1:44] 83 65 136 13 154 118 100 30 205 188 ...
- attr(*, "dimnames")=List of 2
..$ : NULL
..$ : chr [1:44] "data" "igpm" "ipca" "iigpdi" ...
- attr(*, "tsp")= num [1:3] 1994 2011 12
- attr(*, "class")= chr [1:2] "mts" "ts"
> data("Canada")
> plot(Canada,nc=2,xlab="")
> VARselect(Canada,lag.max=8,type="both")
> Canada<-Canada[,c("prod","e","U","rw")]
> pict<-VAR(Canada,p=1,type="both")
> summary(pict,equation="e")
> varcov[1:4,1:4]=summary(pict)$covres[1:4,1:4] #getting the covar . I had
> problem with this comand, so i used instead as below:
>getVarCov=summary(pict)$covres[1:4,1:4]
> c=chol(varcov) #Choleski
> u=rnorm(4) #vetcor with 4 random variables N(0,1)
> P=c%*%u #matricial multiplication choleski%*%rnorm
> xteste=matrix(1:6,nrow=6,ncol=1) # generates a test vector for X
### Start - coefficients of each equation ###
> coef_e=pict$varresult$e$coefficients
> coef_rw=pict$varresult$rw$coefficients
> coef_U=pict$varresult$U$coefficients
> coef_prod=pict$varresult$prod$coefficients
> coef=rbind(coef_prod,coef_e,coef_U,coef_rw)
### End - generates coef matrix with the coefficients of each equation
> coef%*%xx+P #simulates vector Yt+1, conditioned to shocks in vector X and
> shocks already ajusted between the variables by the covar
________________________________
De: David Winsemius <dwinsem...@comcast.net>
Cc: "r-help@r-project.org" <r-help@r-project.org>
Enviadas: Segunda-feira, 17 de Outubro de 2011 22:09
Assunto: Re: [R] Beginner's question about plotting variables in a time series
object with the date on the x axis
On Oct 17, 2011, at 5:23 PM, Iara Faria wrote:
> Dear R helpers,
>
> I am a beginner at R so please be gentle :)
> I have already read manuals and FAQs, with no help.
> I have a monthly time series data on public debt with 40 variables, it starts
> on January 1994 and ends on June 2011.
I agree with Weyland. Figuring out how to use class `ts` objects in R is
difficult. If it were easy I don't think Zeileis would have invented `zoo`.
> I am loading the data into R using read.csv and the data looks ok when I do
> edit(dados):
>> dados<-read.csv("dadosR3.csv", header=T)
> then I tried making it into a time series object, using:
>> dados2<-ts(dados, start = c(1994,1), frequency=12)
(See below for what that did.)
>
> Now when I try plotting any of the variables, for example divliq.pib (net
> debt), the date doesn't appear on the x axis. How can I do this?
> I use
>> plot.ts(divliq.pib)
Try:
plot(time(dados2), dados[, "divliq.pib"] , xy.labels=FALSE))
>
> also the function names() doesn't work, which would be important since I have
> so many variables.
>> names(dados2)
Read the help page again(?). "A data frame will be coerced to a numeric matrix
via data.matrix."
The names function only works for lists and name vectors. Try colnames()
> NULL
>
> I am using R.2.13.1
>
> Thank you very much.
> [[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.
David Winsemius, MD
West Hartford, CT
[[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.
