Skip to content
Create a gist now

Instantly share code, notes, and snippets.

Embed URL


Subversion checkout URL

You can clone with
Download ZIP
Annotated version of myboot
myboot<-function(data, stat, nreps, hist = TRUE) {
# Where 'data' is the original data set to be paste into 'myboot' program.
# 'stat' is the function that will generate the desired statistic such as "standard errors", "confidence intervals" etc.
# 'nreps' is the number of repetitions we want in the simulation.
estimates<-get(stat)(data)# Compute the number of estimates needed:
# Make a container object matrix for store the bootstrap results:
container<-matrix(NA, ncol = len , nrow = nreps)
# The length of "estimates" is the number of coefficients we will estimate standard errors for.
nobs<-nrow(data)#Compute the number of observations to resample
for(i in 1:nreps) {
# Now let's start our bootstrapping resampling loop
# Draw nobs uniform integers draws from 1 to nobs
# This will sample a random sample from the original data
# This apply the function 'stat' to the new data table and store it as a row in 'container' object
### 1.2 ### This will compute standard deviation for each column
if(hist==T) {
frame() # alias for
if(len<= 3) par(mfrow=c(len,1))
if((len> 3)&(len<= 6)) par(mfrow=c(3,2))
for(j in 1:len) hist(container[,j],
main=paste("Estimates for ", names(estimates)[j]), xlab="")
return(list(estimation=container, sds=sds))
###@ 2 @###
model<-lm(lw~s+iq, data=griliches76)
# This function generates a vector with the boostrap estimates.
#mod1<-function(griliches76)lm(lw~s+iq, data=griliches76)[[1]]
mod1.res<-myboot(griliches76, "mod", 10000, hist=T)
# This will run the bootstrap function on the estimates 10.000 times and plot a histogram for the found distribution.

Here's a remix of your code. The resampling is done with sample, and the function will show a progress bar. I have used the foreach package to make the computation parallelizable.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Something went wrong with that request. Please try again.