liesel.model

Liesel modeling framework.

Functions

dist_reg_mcmc(model, seed, num_chains[, ...])	Configures an EngineBuilder for a distributional regression model.
tau2_gibbs_kernel(group)	Builds a Gibbs kernel for a smoothing parameter with an inverse gamma prior.
PIT(_input[, distribution, name])	A weak variable evaluating a probability integral transform (PIT).
Addition(*inputs[, distribution, name])	A weak variable calculating an element-wise sum.
Bijector(_input, bijector[, inverse, ...])	A weak variable evaluating the forward() or inverse() method of a TFP bijector.
ColumnStack(*inputs[, distribution, name])	A weak variable stacking its inputs column-wise.
DesignMatrix(value[, distribution, name])	A strong variable representing a design matrix.
Hyperparameter(value[, distribution, name])	A strong variable representing a hyperparameter.
InverseLink(_input, bijector[, inverse, ...])	A weak variable representing an inverse link function.
Parameter(value[, distribution, name])	A strong variable representing a model parameter.
Predictor(*inputs[, distribution, name])	A weak variable representing a regression predictor.
RegressionCoef(value[, distribution, name])	A strong variable representing a vector of regression coefficients.
Response(value[, distribution, name])	A strong variable representing a response vector.
Smooth(x, beta[, distribution, name])	A weak variable calculating the matrix-vector product x @ beta.
SmoothingParam(value[, distribution, name])	A strong variable representing a smoothing parameter.
load_model(file)	Loads a model from a dill file.
save_model(model, file)
plot_nodes(model[, show, save_path, width, ...])	Plots the nodes of a Liesel model.
plot_vars(model[, show, save_path, width, ...])	Plots the variables of a Liesel model.

Classes

DistRegBuilder()	A model builder for distributional regression models.
GraphBuilder([to_float32])	A graph builder, used to set up a Model.
Model(nodes_and_vars[, grow, copy, to_float32])	A model with a static graph.
Calc(function, *inputs[, _name, ...])	A Node subclass that calculates its value based on its inputs nodes.
Data(value[, _name])	A Node subclass that holds constant data.
Dist(distribution, *inputs[, _name, _needs_seed])	A Node subclass that wraps a probability distribution.
Group(name, **nodes_and_vars)	A group holds a collection of related Var and/or Node objects.
InputGroup(*inputs[, _name, _needs_seed])	A node that groups its inputs for another node.
Node(*inputs[, _name, _needs_seed])	A node of a computational graph that can cache its value.
NodeState(value, outdated[, extra])	The state of a node.
TransientCalc(function, *inputs[, _name, ...])	A transient calculator node that does not cache its value.
TransientDist(distribution, *inputs[, ...])	A transient distribution node that does not cache its value.
TransientIdentity(_input[, _name])	A transient identity node that does not cache its value.
TransientNode(*inputs[, _name, _needs_seed])	A node that does not cache its value.
Value(value[, _name])	A Node subclass that holds constant values.
Var(value[, distribution, name, inference])	A variable in a statistical model.