Continuous-Time Bayesian Networks
A CTBN(Continuous-Time Bayesian Networks) is a graphical model that allows a Bayesian Network to evolve over continuous time. This pyAgrum library offers ways to create such models, to have a graphical representation but also to learn such models (the dependency between variables and their distribution parameters) using exact inference and sampling as well. To this day Forward Sampling is the only sampling method available.
The goal is to have the properties of a discrete Markov Chain but at continuous time, which means that a random variable is allowed to switch state at any time. To depict the time a variable spends in a state before switching to another, we use an exponential distribution.
Tutorials
Reference
- The CTBN model
CIM
CIM.DELIMITER
CIM.add()
CIM.amalgamate()
CIM.asPotential()
CIM.extract()
CIM.findVar()
CIM.fromMatrix()
CIM.getPotential()
CIM.instantiation()
CIM.isIM()
CIM.isParent()
CIM.nbrDim()
CIM.remove()
CIM.toMatrix()
CIM.varI()
CIM.varJ()
CIM.varNames
CIM.varRadical()
CIM.variable()
CIM.variablesSequence()
CTBN
CTBN._graph
CTBN._cim
CTBN._id2var
CTBN._name2id
CTBN.CIM()
CTBN.add()
CTBN.addArc()
CTBN.arcs()
CTBN.children()
CTBN.childrenNames()
CTBN.completeInstantiation()
CTBN.equals()
CTBN.eraseArc()
CTBN.fullInstantiation()
CTBN.labels()
CTBN.name()
CTBN.names()
CTBN.node()
CTBN.nodes()
CTBN.parentNames()
CTBN.parents()
CTBN.toDot()
CTBN.variable()
CTBN.variables()
- Other functions
- Inference
- Graphical tools
- Learning a CTBN