penalized_zero_inflated_truncated_gaussian
Regularized zero-inflated truncated Gaussian marginal model.
PenalizedZeroInflatedTruncatedGaussian(formula, *, prior_alpha=1.1, penalties=(), device=None)
A regularized zero-inflated truncated Gaussian marginal
The penalty P is added to the training loss, together with a
Beta(prior_alpha, prior_alpha) prior term already present in
:class:ZeroInflatedTruncatedGaussian.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
formula
|
dict of str to str, or str
|
|
required |
prior_alpha
|
float
|
See :class: |
1.1
|
penalties
|
sequence of CoefficientPenalty
|
See :mod: |
()
|
device
|
device or str
|
|
None
|
Examples:
>>> from scdesigner.distributions.penalties import CoefficientPenalty, QuadraticPenalty
>>> penalty = CoefficientPenalty("mean", ["x"], [1.0], QuadraticPenalty(lam=0.1))
>>> model = PenalizedZeroInflatedTruncatedGaussian({"mean": "~ x"}, penalties=[penalty])
>>> type(model).__name__
'ZeroInflatedTruncatedGaussian'
Source code in scdesigner/src/scdesigner/distributions/penalized_zero_inflated_truncated_gaussian.py
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 | |