Tutorials on pyAgrum

Tutorial pyAgrum

Using pyAgrum

Exact and Approximated Inference

Probablistic Inference with pyAgrum

Relevance Reasoning with pyAgrum

Some other features in Bayesian inference

Approximate inference in aGrUM (pyAgrum)

Different sampling inference

Learning Bayesian networks

Learning the structure of a Bayesian network

Learning BN as probabilistic classifier

Learning essential graphs

Dirichlet prior

Parametric EM (missing data)

Scores, Chi2, etc. with BNLearner

Different Graphical Models

Influence diagram

Dynamic Bayesian Networks

Markov random fields (a.k.a. Markov Networks)

Credal Networks

Object-Oriented Probabilistic Relational Model

Bayesian networks as scikit-learn compliant classifiers

Learning classifiers

Discretization using pyAgrum’s DiscreteTypeProcessor

Comparing classifiers (including Bayesian networks) with scikit-learn

Using sklearn to cross-validate bayesian network classifier

From a Bayesian network to a Classifier

Causal Bayesian Networks

Smoking, Cancer and causality

Simpson’s Paradox

Multinomial Simpson Paradox

Some examples of do-calculus

Counterfactual : the Effect of Education and Experience on Salary

Causal Effect Estimation in datasets

pyAgrum’s (experimental) models

Using Continuous-Time Bayesian Networks

Conditional Linear Gaussian models

Bayesian Network Mixture (gbnm) model

Quantum Bayesian Network Sampling (pyagrum.qBNSampling)

pyAgrum’s specific features

Loading and saving graphical models

SHAP values, SHALL values

Named tensors

Aggregators

Explaining a model

Kullback-Leibler for Bayesian networks

Comparing BNs

Customizing and exporting graphical models and CPTs as image (pdf, png)

gum.config :the configuration object for pyAgrum

Examples

Kaggle Titanic

Naive modeling of credit defaults using a Markov Random Field

Learning and causality

Sensitivity analysis for Bayesian networks using credal networks

Quasi-continuous BN

Bayesian Beta Distributed Coin Inference

ACE estimations from real observational data

interactive notebooks

Examples from Pearl and Mackenzie [PM18]

MiniTuring (p46)

Smallpox Paradox (p50)

Where is my Bag ? (p115)

Walking Example (p135)

Back-Door Criterion (p150)

Smoking (chapter 5)

Monty Hall Problem (p178)

Do-Calculus (p213)

The Curious Case(s) For Dr. Snow (p224)

Good and Bas Cholesterol (p229)

The Effect of Education and Experience on Salary (p251)

[PM18]

Judea Pearl and Dana Mackenzie. The Book of Why: The New Science of Cause and Effect. Basic Books, 2018. ISBN 978-0-465-09760-9.