Stixbox Function
Last update : 21/4/2006

ciboot - Various bootstrap confidence intervals

Calling Sequence

ci = ciboot(x,'T',method,C,B)

Description

Compute a bootstrap confidence interval for T(X) with level C. Default for C is 0.90. Number of resamples is B, with default that is different for different methods. The method is

1: Normal approximation (std is bootstrap).
2 : Simple bootstrap principle (bad, don't use).
3 : Studentized, std is computed via jackknife (If T is 'mean' this done the fast way via the routine TEST1B).
4 : Studentized, std is 30 samples' bootstrap.
5 : Efron's percentile method.
6: Efron's percentile method with bias correction (BC).

Default method is 5. Often T(X) is a number but it may also be a vector or a even a matrix. Every row of the result ci is of the form

       [LeftLimit, PointEstimate, RightLimit]
   

and the corresponding element of T(X) is found by noting that t = T(X); t = t(:); is used in the routine.

Examples

X = rand(13,2), C = cov(X), ci = ciboot(X,'cov')
 

See Also

stdboot,  stdjack,