HSEL.bigz: Sample without replacement when the population can be huge.

In MRPP: Multiresponse permutation procedure and its variable importance and variable selection methods.

Description

General a small sample of size M out of a huge population of size N.

Usage

HSEL.bigz(N, M, seed=0)
rejsample(N, M, seed=0)
sample1(N, seed=0)

Arguments

N	The size of the population.
M	The sample size.
seed	A bigz integer of random number seed. See urand.bigz.

Details

HSEL.bigz generates a sample of size M from a population of size N, with space complexity linear in M. rejsample is similar to HSEL.bigz, but implemented using direct rejection through unique. sample1 generates one number of 0 to N-1 using simple rejection and urand.bigz.

Value

For HSEL.bigz and rejsample, a bigz vector of length N, between 1 and N. For sample1, a bigz integer, between 0 and N-1.

Note

rejsample can be very slow when M is large.

References

Jarmo Ernvall and Olli Nevalainen. 1982. An Algorithm for Unbiased Random Sampling. THE COMPUTER JOURNAL, VOL. 25, NO. 1, 45–47

See Also

sample, urand.bigz

Examples

HSEL.bigz(factorialZ(15),5,12345)
rejsample(factorialZ(15),5,12345)
sample1(factorialZ(15), 12345)

MRPP documentation built on May 2, 2019, 4:46 p.m.