This small package implements methods for sampling from a unit simplex, a problem that often crops up in a data analysis context.
There is only a single sampling strategy that results in uniform samples from the unit simplex:
from simsam import kraemer_sampling
# Sample 1,000 points from the 10-dimensional unit simplex.
dim = 10
N = 1000
samples = kraemer_sampling(dim, N)For comparison purposes, there is also a naive sampling procedure, which does not result in uniform samples.
from simsam import naive_sampling
# Sample 1,000 points from the 10-dimensional unit simplex. Notice that
# the samples will be biased.
dim = 10
N = 1000
samples = naive_sampling(dim, N)