bnlearn is Python package for causal discovery by learning the graphical structure of Bayesian networks, parameter learning, inference, and sampling methods.
Because probabilistic graphical models can be difficult to use, Bnlearn contains the most-wanted pipelines. Navigate to API documentations for more detailed information. ⭐️ Star it if you like it ⭐️
| Feature | Description | Medium | Gumroad/Podcast |
|---|---|---|---|
| Causal Discovery - Overview and Starters Guide | Learn the basics of causal modelling. | link | --- |
| Structure Learning | Learn the model structure from data or with expert knowledge. | link | Podcast |
| Causal Predictions | Learn to make causal predictions and explore interventions. | link | --- |
| Parameter Learning | Estimate model parameters (e.g., conditional probability distributions) from observed data and combine data with expert knowledge. | link | Podcast |
| Causal Inference | Compute interventional and counterfactual distributions using Bayesian inference. | link | Podcast |
| Generate Synthetic Data | Generate synthetic data using Bayesian sampling. | link | Podcast |
| Discretize Data | Discretize continuous datasets for Bayesian network modelling. | --- | --- |
| Data Types | Discrete, continuous and mixed datasets. | --- | --- |
| Comparisons | Compare causal discovery libraries and their ability to find hidden causal relationships. | link | Podcast |
| LLMs and Causal Discovery | Benchmark large language models against Bayesian methods for causal discovery. | link | --- |
- Example Notebooks: Examples
- Blog Posts: Medium
- Documentation: Website
- Bug Reports and Feature Requests: GitHub Issues
| Feature | Description |
|---|---|
| Key Pipelines | |
| Structure learning | bn.structure_learning.fit() |
| Parameter learning | bn.parameter_learning.fit() |
| Inference | bn.inference.fit() |
| Make predictions | bn.predict() |
| Generate Synthetic Data | bn.sampling() |
| Compute Edge Strength | bn.independence_test() |
| Key Functions | |
| Imputations | bn.knn_imputer() |
| Discretizing | bn.discretize() |
| Check Model Parameters | bn.check_model() |
| Create DAG | bn.make_DAG() |
| Get Node Properties | bn.get_node_properties() |
| Get Edge Properties | bn.get_edge_properties() |
| Get Parents From Edges | bn.get_parents() |
| Generate Default CPT per Node | bn.generate_cpt() |
| Generate Default CPTs for All Edges | bn.build_cpts_from_structure() |
| Make Plots | |
| Plotting | bn.plot() |
| Plot Graphviz | bn.plot_graphviz() |
| Compare 2 Networks | bn.compare_networks() |
| Load DAG (bif files) | bn.import_DAG() |
| Load Examples | bn.import_example() |
| Transformation Functions | |
| Convert DAG to Undirected | bn.to_undirected() |
| Convert to one-hot | bn.df2onehot() |
| Convert Adjacency Matrix to Vector | bn.adjmat2vec() |
| Convert Adjacency Matrix to Dictionary | bn.adjmat2dict() |
| Convert Vector to Adjacency Matrix | bn.vec2adjmat() |
| Convert DAG to Adjacency Matrix | bn.dag2adjmat() |
| Convert DataFrame to Onehot | bn.df2onehot() |
| Convert Query to DataFrame | bn.query2df() |
| Convert Vector to DataFrame | bn.vec2df() |
| Metrics | |
| Compute Topological Ordering | bn.topological_sort() |
| Compute Structure Scores | bn.structure_scores() |
| General | |
| Save Model | bn.save() |
| Load Model | bn.load() |
| Print CPTs | bn.print_CPD() |
Python 3.10 or newer is required.
uv venv
.venv\Scripts\activate
uv pip install bnlearnuv venv
.venv\Scripts\activate
git clone https://github.com/erdogant/bnlearn.git
cd bnlearn
uv pip install -U -e .conda create -n env_bnlearn python=3.12
conda activate env_bnlearn
pip install bnlearnpip install git+https://github.com/erdogant/bnlearn
uv pip install git+https://github.com/erdogant/bnlearnuv pip install -e ".[dev]"
uv run pytestpip install -e ".[dev]"
pytest -vnpx skills add erdogant/bnlearn --skill bnlearn# Import library
import bnlearn as bn import bnlearn as bn
# Example dataframe sprinkler_data.csv can be loaded with:
df = bn.import_example()
# df = pd.read_csv('sprinkler_data.csv')
Cloudy Sprinkler Rain Wet_Grass
0 0 1 0 1
1 1 1 1 1
2 1 0 1 1
3 0 0 1 1
4 1 0 1 1
.. ... ... ... ...
995 0 0 0 0
996 1 0 0 0
997 0 0 1 0
998 1 1 0 1
999 1 0 1 1
model = bn.structure_learning.fit(df)
# Compute edge strength with the chi-square test statistic
model = bn.independence_test(model, df)
G = bn.plot(model)
# Example: Structure Learning
model_hc_bic = bn.structure_learning.fit(df, methodtype='hc', scoretype='bic')
model_hc_k2 = bn.structure_learning.fit(df, methodtype='hc', scoretype='k2')
model_hc_bdeu = bn.structure_learning.fit(df, methodtype='hc', scoretype='bdeu')
model_ex_bic = bn.structure_learning.fit(df, methodtype='ex', scoretype='bic')
model_ex_k2 = bn.structure_learning.fit(df, methodtype='ex', scoretype='k2')
model_ex_bdeu = bn.structure_learning.fit(df, methodtype='ex', scoretype='bdeu')
model_cl = bn.structure_learning.fit(df, methodtype='cl', root_node='Wet_Grass')
model_tan = bn.structure_learning.fit(df, methodtype='tan', root_node='Wet_Grass', class_node='Rain')
# Example: Parameter Learning
import bnlearn as bn
# Import dataframe
df = bn.import_example()
# As an example we set the CPD at False which returns an "empty" DAG
model = bn.import_DAG('sprinkler', CPD=False)
# Now we learn the parameters of the DAG using the df
model_update = bn.parameter_learning.fit(model, df)
# Make plot
G = bn.plot(model_update)
# Example: Inference
import bnlearn as bn
model = bn.import_DAG('sprinkler')
query = bn.inference.fit(model, variables=['Rain'], evidence={'Cloudy':1,'Sprinkler':0, 'Wet_Grass':1})
print(query)
print(query.df)
# Lets try another inference
query = bn.inference.fit(model, variables=['Rain'], evidence={'Cloudy':1})
print(query)
print(query.df)Thank the contributors!
- Erdogan Taskesen, github: erdogant
- Contributions are welcome.
- Yes! This library is entirely free but it runs on coffee! :) Feel free to support with a Coffee.
