[R] Seeking help with code
Stephanie Tsalwa
stsalwa at hotmail.com
Thu Dec 7 23:51:55 CET 2017
Assuming the days of raining during half a year of all states(provinces) of a country is normally distributed (mean=ì, standard deviation=ó) with sigma (ó) equals to 2. We now have 10 data points here: 26.64, 30.65, 31.27, 33.04, 32.56, 29.10, 28.96, 26.44, 27.76, 32.27. Try to get the 95% level of CI for ì, using parametric Bootstrap method with bootstrap size B=8000.
my code - what am i doing wrong
#set sample size n, bootstrap size B
n = 10
b = 8000
set a vector of days of rain into "drain"
drain = c(26.64, 30.65, 31.27, 33.04, 32.56, 29.10, 28.96, 26.44, 27.76, 32.27)
#calculate mean of the sample for days of rain
mdr=mean(drain)
mdr
#calculate the parameter of the exponential distribution
lambdahat = 1.0/mdr
lambdahat
#draw the bootstrap sample from Exponential
x = rexp(n*b, lambdahat)
x
bootstrapsample = matrix(x, nrow=n, ncol=b)
bootstrapsample
# Compute the bootstrap lambdastar
lambdastar = 1.0/colMeans(bootstrapsample)
lambdastar
# Compute the differences
deltastar = lambdastar - lambdahat
deltastar
# Find the 0.05 and 0.95 quantile for deltastar
d = quantile(deltastar, c(0.05,0.95))
d
# Calculate the 95% confidence interval for lambda.
ci = lambdahat - c(d[2], d[1])
ci
[[alternative HTML version deleted]]
More information about the R-help
mailing list