Quickstart

Installation (short)

Installation of bnlearn is straightforward. It is advisable to create a new environment.

conda create -n env_bnlearn python=3.8
conda activate env_bnlearn
pip install bnlearn

Quick Examples

Let’s start by importing some data. We need a DAG and CPD.

import bnlearn as bn

df = bn.import_example()

model = bn.structure_learning.fit(df)

model = bn.independence_test(model, df)

G = bn.plot(model)

import bnlearn as bn

model = bn.import_DAG('sprinkler')

df = bn.import_example()

df = bn.sampling(model)

query = bn.inference.fit(model, variables=['Rain'], evidence={'Cloudy':1, 'Wet_Grass':1})
print(query.df)

model_sl = bn.structure_learning.fit(df)

model_pl = bn.parameter_learning.fit(model_sl, df)

scores, adjmat = bn.compare_networks(model_sl, model)

---

---