Use R programming for "As was described in the view graphs, build bootstrap calculator and simulation calculator
Calculate for sample sizes of 3, 10,30, and 100, and alpha =.1, .05
Create a table of the results in word, or a text document and write up your conclusions based on the table, send me the document, and text copies or a dump file of your r programs."
- Build a confidence interval estimator for the mean based on the bootstrap (use 10000 =nboot)
- Build a simulator that draws n samples form a lognormal distribution (rlnorm) and builds both the central limit theorem based confidence interval, and compares it to the coverage rate for the bootstrap (confidence interval based on the 1st program). (1000 simulation runs minimum)
- Compare the coverage rates for the bootstrap confidence interval and the central limit theorem based confidence interval.
For sample sizes 5, 10,15, and 30
- Now build the bootstrap estimator into a simulator which compares how well the bootstrap confidence interval covers the true value vs a normal theory confidence interval.
- Run it for lognormal (nsim=1000) for n=3,10,30,100, alpha= .1,.05 That's 8 runs
- Compare the coverage of the 2 confidence intervals with nominal coverage
- Write up the results with a table