For a hypergeometric distribution, denoted by Hyper(n_pop, k_s, n_draw), where n_pop is the population size, k_s is the number of population with 'success', and n_draw is the given sample size, this program calculates the confidence interval for k_s when n_pop is known and number of observed success is given as k_s_obs, using simulation scipy optimization tools.
Note that this is done by a simulation and can have unaccurate values as a solution. Please cross-check with other site calculating confidence interval of hypergeometric function like here: cluster-text.com.
PyPI package link is here.
First, pip install this package.
shell
pip install cisimNote: if this installing doesn't go well, please try pip install these packages: cerberus, numpy, and scipy.
After the pip-install, you can calculate the interval of a Hypergeometric distribution as follows.
python
import cisim
from cisim.stats import HyperCI
h = HyperCI(n_pop=10**4, n_draw=10**3, k_s_obs=100)
res = h.ci_sim()
print(res[0])
>>> [830, 1193]And this program can also calculate confidence interaval for binomial distribution, too.
import cisim
from cisim.stats import BinomCI
b = BinomCI(n_pop=100, n_obs=10, cl=0.05)
res = b.ci_sim()
print(res[0])
>>> [0.049005430267763495, 0.17622473596973592]You can check and compare the result with Clopper-Pearson approach on cluster-text.com
Learn more at HyperGeometric distribution, Binomial Distribution
If you want to learn more about setup.py files, check out this repository.