Open robmoss opened 4 days ago
This modifies the lhs_prcc function so that it can accept a dictionary that maps parameter names to percentage-point (inverse-CDF) functions. Here is an example of what it supports:
dist_nu = scipy.stats.beta(a=1, b=1, loc=0.767, scale=0.025) dist_nu_lag = scipy.stats.beta(a=100, b=100, loc=1 / 28, scale=1 / 7 - 1 / 28) dist_l_v = scipy.stats.beta(a=100, b=0.5, loc=0.3, scale=0.3) dist_l_h = scipy.stats.beta(a=0.5, b=100, loc=0.7, scale=0.25) parameter_dists = { 'nu_unvaccinated': dist_nu.ppf, 'nu_vaccination_lag': dist_nu_lag.ppf, 'l_v': dist_l_v.ppf, 'h_v': dist_l_h.ppf, } results_df, prccs = lhs_prcc( parameters_df=parameter_dists, # other arguments ... )
I've added a new test case in tests/test_lhs_arbitrary_dists.py to check that the samples are consistent with the defined distributions.
tests/test_lhs_arbitrary_dists.py
This modifies the lhs_prcc function so that it can accept a dictionary that maps parameter names to percentage-point (inverse-CDF) functions. Here is an example of what it supports:
I've added a new test case in
tests/test_lhs_arbitrary_dists.py
to check that the samples are consistent with the defined distributions.