pyAgrum 1.14.0

Bayesian networks and other Probabilistic Graphical Models.

pyAgrum

pyAgrum is a scientific C++ and Python library dedicated to Bayesian Networks and other Probabilistic Graphical Models. It provides a high-level interface to the part of aGrUM allowing to create, model, learn, use, calculate with and embed Bayesian Networks and other graphical models. Some specific (python and C++) codes are added in order to simplify and extend the aGrUM API.

Example

import pyAgrum as gum

# Creating BayesNet with 4 variables
bn=gum.BayesNet('WaterSprinkler')
print(bn)

# Adding nodes the long way
c=bn.add(gum.LabelizedVariable('c','cloudy ?',["Yes","No"]))
print(c)

# Adding nodes the short way
s, r, w = [ bn.add(name, 2) for name in "srw" ]
print (s,r,w)
print (bn)

# Addings arcs c -> s, c -> r, s -> w, r -> w
bn.addArc(c,s)
for link in [(c,r),(s,w),(r,w)]:
bn.addArc(*link)
print(bn)

# or, equivalenlty, creating the BN with 4 variables, and the arcs in one line
bn=gum.fastBN("w<-r<-c{Yes|No}->s->w")

# Filling CPTs
bn.cpt("c").fillWith([0.5,0.5])
bn.cpt("s")[0,:]=0.5 # equivalent to [0.5,0.5]
bn.cpt("s")[{"c":1}]=[0.9,0.1]
bn.cpt("w")[0,0,:] = [1, 0] # r=0,s=0
bn.cpt("w")[0,1,:] = [0.1, 0.9] # r=0,s=1
bn.cpt("w")[{"r":1,"s":0}] = [0.1, 0.9] # r=1,s=0
bn.cpt("w")[1,1,:] = [0.01, 0.99] # r=1,s=1
bn.cpt("r")[{"c":0}]=[0.8,0.2]
bn.cpt("r")[{"c":1}]=[0.2,0.8]

# Saving BN as a BIF file
gum.saveBN(bn,"WaterSprinkler.bif")

# Loading BN from a BIF file
bn2=gum.loadBN("WaterSprinkler.bif")

# Inference
ie=gum.LazyPropagation(bn)
ie.makeInference()
print (ie.posterior("w"))

# Adding hard evidence
ie.setEvidence({"s": 1, "c": 0})
ie.makeInference()
print(ie.posterior("w"))

# Adding soft and hard evidence
ie.setEvidence({"s": [0.5, 1], "c": 0})
ie.makeInference()
print(ie.posterior("w"))

LICENSE

Copyright (C) 2005-2024 by Pierre-Henri WUILLEMIN et Christophe GONZALES {prenom.nom}_at_lip6.fr

The aGrUM/pyAgrum library and all its derivatives are distributed under the LGPL3 license, see https://www.gnu.org/licenses/lgpl-3.0.en.html.

Authors

• Pierre-Henri Wuillemin

• Christophe Gonzales

Maintainers

• Lionel Torti

• Gaspard Ducamp