import numpy as np
[docs]
def combine_samples(baseline, perturb):
"""Combine baseline and perturbation samples in order.
The number of parameters will be inferred from the shape of `baseline`.
Arguments
---------
baseline : np.array
baseline sample for one of `N` samples
perturb : np.array
perturbation value for a parameter
Example
-------
>>> X = combine_samples(baseline, perturb)
`X` will now hold:
[[x_{1,1}, x_{1,2}, ..., x_{1,p}]
[b_{1,1}, x_{1,2}, ..., x_{1,p}]
[x_{1,1}, b_{1,2}, ..., x_{1,p}]
[x_{1,1}, x_{1,2}, ..., b_{1,p}]
...
[x_{N,1}, x_{N,2}, ..., x_{N,p}]
[b_{N,1}, x_{N,2}, ..., x_{N,p}]
[x_{N,1}, b_{N,2}, ..., x_{N,p}]
[x_{N,1}, x_{N,2}, ..., b_{N,p}]]
where `p` denotes the number of parameters as specified in
`problem` and `N` is the number of samples.
We can now run the model using the values in `X`.
The total number of model evaluations will be `N(p+1)`.
Returns
-------
np.array
"""
assert (
baseline.shape == perturb.shape
), "Baseline and perturbation data should be of same size"
nrows, num_vars = baseline.shape
group = num_vars + 1
sample_set = np.repeat(baseline, repeats=group, axis=0)
grp_start = 0
for i in range(nrows):
mod = np.diag(perturb[i])
first_ = grp_start + 1
np.copyto(sample_set[first_ : first_ + num_vars], mod, where=mod != 0.0)
grp_start += group
# End for
return sample_set
