funclp.Distribution module

class funclp.Distribution(**kwargs)[source]

Bases: ABC, CudaReference

Class defining the noise distribution in data.

Parameters:

kwargs (dict) – Attributes to change.

pdf[source]

Probability Density Function for the given distribution

Type:

method

loglikelihood_reduced[source]

Log-likelihood up to additive constants.

Type:

method

loglikelihood[source]

Exact log-likelihood (with constants).

Type:

method

dloglikelihood[source]

Derivative of log-likelihood w.r.t model parameter.

Type:

method

d2loglikelihood[source]

Second derivative of log-likelihood (observed curvature).

Type:

method

fisher[source]

Expected curvature (Fisher information).

Type:

method

abstractmethod d2loglikelihood(model_data, weights=np.float32(1.0), /, ignore=np.False_)[source]
abstract property default_attributes
abstractmethod dloglikelihood(model_data, weights=np.float32(1.0), /, ignore=np.False_)[source]
abstractmethod fisher(model_data, weights=np.float32(1.0), /, ignore=np.False_)[source]
abstractmethod loglikelihood(model_data, weights=np.float32(1.0), /, ignore=np.False_)[source]
abstractmethod loglikelihood_reduced(model_data, weights=np.float32(1.0), /, ignore=np.False_)[source]
property name
abstractmethod pdf(model_data, weights=np.float32(1.0), /, ignore=np.False_)[source]

Distributions