[R] calculate a "rolling mean" including standard deviation

[Rui Barradas]

Hello,

Something like this?


d <- myframe2$distance
n <- length(d)
mu <- rep(NA, n - 1)
mu[n - 1] <- m <- mean(d[(n - 1):n])
s <- sd(d[(n - 1):n])
ex <- 0
for(i in rev(seq_len(n))[-(1:2)]){
     if(d[i] < m + s){
         ex <- 0
         mu[i] <- m
     }else{
         ex <- ex + 1
         if(ex >= 2) break
     }
     m <- mean(d[i:n])
     s <- sd(d[i:n])
}
mu


Hope this helps,

Rui Barradas
Em 17-12-2012 14:55, Tagmarie escreveu:
> Hello everyone,
>
> I have a data frame somewhat like this one:
>
> myframe <- data.frame (Timestamp=c( "24.09.2012 06:00", "24.09.2012 07:00",
> "24.09.2012 08:00",
>                                      "24.09.2012 09:00", "24.09.2012 10:00",
> "24.09.2012 11:00",
>                                      "24.09.2012 12:00", "24.09.2012 13:00",
> "24.09.2012 14:00"),
>                          distance =c(9,9,9,4,5,9,4,5,5 ) )
> myframestime <- as.POSIXct (strptime(as.character(myframe$Timestamp),
> "%d.%m.%Y %H:%M"), tz="GMT")
> myframe2 <- cbind (myframestime, myframe)
> myframe2$Timestamp <- NULL
> myframe2
>
> This is what I want to do:
> 1.) calculate the mean and the standard deviation for "distance" from the
> last too rows (at 13:00 and 14:00)
> 2.) compare the value for distance one row earlier (12:00). If that value is
> in the range of the previously calculated mean + sd I want to include the
> value and calculate a new mean and a new sd. If there is one value which is
> not in the range I want to exclude/ignore the value. If there are two
> subsequent values which are not in the range then I want to stopp the
> calculation (or at least mark the point by including e.g. NAs).
>
> Does anyone know how to do that?
>
>
>
>
>
> --
> View this message in context: http://r.789695.n4.nabble.com/calculate-a-rolling-mean-including-standard-deviation-tp4653303.html
> Sent from the R help mailing list archive at Nabble.com.
>
> ______________________________________________
> R-help at 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.