[R] extracting bootstrap statistics by group with loop

[Marna Wagley]

Thank you Rui. It helped a lot.
-MW

On Sun, Oct 31, 2021 at 10:16 AM Rui Barradas <ruipbarradas using sapo.pt> wrote:

> Hello,
>
> Now I'm spamming the list, not one of my days.
>
> My first post was right, there was no bug and the 2nd one was exactly
> the same code, it corrected nothing at all.
>
> Apologies for the noise,
>
> Rui Barradas
>
> Às 16:55 de 31/10/21, Rui Barradas escreveu:
> > Hello,
> >
> > Sorry, bug. In both by instructions it's boot_mean_se, not bootprop.
> >
> >
> > boot_year <- by(DaT, DaT$Year, boot_mean_se, statistic = bootprop, R = R)
> > boot_year_area <- by(DaT,
> >                       INDICES = list(Year = DaT$Year, Area = DaT$Area),
> >                       FUN = boot_mean_se,
> >                       statistic = bootprop, R = R)
> >
> >
> > Hope this helps,
> >
> > Rui Barradas
> >
> > Às 16:48 de 31/10/21, Rui Barradas escreveu:
> >> Hello,
> >>
> >> Try to aggregate with ?by.
> >>
> >>
> >> bootprop <- function(data, index){
> >>    d <- data[index, ]
> >>    sum(d[["bothTimes"]], na.rm = TRUE)/sum(d[["total"]], na.rm = TRUE)#
> >> }
> >> boot_mean_se <- function(data, statistic, R){
> >>    b <- boot::boot(DaT, bootprop, R = R)
> >>    c(bootMean = mean(b$t), bootSE = sd(b$t))
> >> }
> >>
> >> boot_year <- by(DaT, DaT$Year, boot_mean_se, statistic = bootprop, R =
> R)
> >> boot_year_area <- by(DaT,
> >>                       INDICES = list(Year = DaT$Year, Area = DaT$Area),
> >>                       FUN = boot_mean_se,
> >>                       statistic = bootprop, R = R)
> >> boot_year
> >> boot_year_area
> >>
> >> boot_year <- do.call(rbind, boot_year)
> >>
> >> d <- dimnames(boot_year_area)
> >> boot_year_area <- cbind(Reduce(expand.grid, rev(d))[2:1],
> >>                          do.call(rbind, boot_year_area))
> >> names(boot_year_area)[1:2] <- names(d)
> >> boot_year_area
> >>
> >>
> >> Hope this helps,
> >>
> >> Rui Barradas
> >>
> >> Às 11:47 de 31/10/21, Marna Wagley escreveu:
> >>> Hi R users,
> >>> I was trying to extract the bootstrap mean and its SE by group but I
> >>> have
> >>> been doing  it by separating the group manually. The data set is big so
> >>> doing it manually is a kind of tedious task. I am wondering whether
> >>> there
> >>> is a possibility to do it by creating a loop. I am weak in writing loop
> >>> functions. I am attaching an example data and how I performed the
> >>> analysis, see below.
> >>> Thanks for your help.
> >>> Sincerely,
> >>> MW
> >>> ####
> >>> library(boot)
> >>> DaT<-structure(list(bothTimes = c(0L, 0L, 1L, 0L, 1L, 0L, 0L, 0L,
> >>> 1L, 0L, 0L, 0L, 1L, 1L, 0L, 1L, 0L, 0L, 1L, 0L, 0L, 1L, 1L, 0L),
> >>> total = c(1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L,
> >>> 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L), Area = c("A",
> >>> "A", "A", "A", "A", "A", "B", "B", "B", "B", "B", "B", "A", "A",
> >>> "A", "A", "A", "A", "B", "B", "B", "B", "B", "B"), Year = c(2015L,
> >>> 2015L,
> >>> 2015L, 2015L, 2015L, 2015L, 2015L, 2015L, 2015L, 2015L,
> >>> 2015L, 2015L, 2016L, 2016L, 2016L, 2016L, 2016L, 2016L, 2016L,
> >>> 2016L, 2016L, 2016L, 2016L, 2016L)), class = "data.frame", row.names =
> >>> c(NA, -24L))
> >>>
> >>> head(DaT)
> >>> R=100
> >>> bootprop <- function(data, index){
> >>>    d <- data[index, ]
> >>>    sum(d[["bothTimes"]], na.rm = TRUE)/sum(d[["total"]], na.rm = TRUE)#
> >>> }
> >>>
> >>> ###################
> >>> #2015
> >>> ###################
> >>> #-----Year2015_pooled
> >>> Y2015_pooled<-subset(DaT, DaT$Year=="2015")
> >>> Y2015_pooled_boot <- boot(Y2015_pooled, bootprop, R)
> >>> boot_Y2015_pooled<-data.frame(Year="2015", Area= "Pooled", bootMean=
> >>> Y2015_pooled_boot$t0, SE=sd(Y2015_pooled_boot$t))
> >>> #-----Year2015_AreaA
> >>> Y2015_A<-subset(DaT, DaT$Year=="2015" & DaT$Area=="A")
> >>> Y2015_A_boot <- boot(Y2015_A, bootprop, R)
> >>> boot_Y2015_A<-data.frame(Year="2015", Area= "A", bootMean=
> >>> Y2015_A_boot$t0,
> >>> SE=sd(Y2015_A_boot$t))
> >>> #----Year2015_AreaB
> >>> Y2015_B<-subset(DaT, DaT$Year=="2015" & DaT$Area=="B")
> >>> Y2015_B_boot <- boot(Y2015_B, bootprop, R)
> >>> boot_Y2015_B<-data.frame(Year="2015", Area= "B", bootMean=
> >>> Y2015_B_boot$t0,
> >>> SE=sd(Y2015_B_boot$t))
> >>> ###################
> >>> #2016
> >>> ###################
> >>> #-----Year2016_pooled
> >>> Y2016_pooled<-subset(DaT, DaT$Year=="2016")
> >>> Y2016_pooled_boot <- boot(Y2016_pooled, bootprop, R)
> >>> boot_Y2016_pooled<-data.frame(Year="2016", Area= "Pooled", bootMean=
> >>> Y2016_pooled_boot$t0, SE=sd(Y2016_pooled_boot$t))
> >>>
> >>> #-----Year2016_AreaA
> >>> Y2016_A<-subset(DaT, DaT$Year=="2016" & DaT$Area=="A")
> >>> Y2016_A_boot <- boot(Y2016_A, bootprop, R)
> >>>
> >>> boot_Y2016_A<-data.frame(Year="2016", Area= "A", bootMean=
> >>> Y2016_A_boot$t0,
> >>> SE=sd(Y2016_A_boot$t))
> >>> #----Year2016_AreaB
> >>> Y2016_B<-subset(DaT, DaT$Year=="2016" & DaT$Area=="B")
> >>> Y2016_B_boot <- boot(Y2016_B, bootprop, R)
> >>> boot_Y2016_B<-data.frame(Year="2016", Area= "B", bootMean=
> >>> Y2016_B_boot$t0,
> >>> SE=sd(Y2016_B_boot$t))
> >>>
> >>> ## output data.matrix
> >>>
> BootMean_All<-rbind(boot_Y2015_pooled,boot_Y2015_A,boot_Y2015_B,boot_Y2016_pooled,boot_Y2016_A,boot_Y2016_B)
>
> >>>
> >>> BootMean_All
> >>>
> >>>     [[alternative HTML version deleted]]
> >>>
> >>> ______________________________________________
> >>> R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
> >>> 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.
> >>>
> >>
> >> ______________________________________________
> >> R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
> >> 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.
> >
> > ______________________________________________
> > R-help using r-project.org mailing list -- To UNSUBSCRIBE and more, see
> > 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.
>

	[[alternative HTML version deleted]]