What is a Bayesian network?

A Bayesian network is a probabilistic graphical model that represents variables and their conditional dependencies using a directed acyclic graph (DAG). Each node has a conditional probability table (CPT) that quantifies the effect of its parents.

Can I run inference with evidence?

Yes. After defining your nodes and CPTs, select a query variable and set evidence for other variables. The tool will perform exact enumeration to compute P(Query | Evidence).

Is this for production?

No. This is an educational, in-browser tool for small Bayesian networks. For large models and performance, use a dedicated library or engine.

Bayesian Network (Interactive)

Create a small directed acyclic graph (DAG) of binary variables, specify conditional probability tables, and run exact inference with evidence — right in the browser.

Note: this tool is for learning. For large models use a specialized engine.

1. Nodes

All nodes are boolean: True / False.

3. Inference

Choose a query variable and (optional) evidence. We'll compute P(Query = True | Evidence).

Query variable

Query value

Evidence

Result:

—

Bayesian network recap

A Bayesian network is a compact way to encode a joint probability distribution using conditional independencies. Formally, for nodes \( X_1, \dots, X_n \) in a DAG, the joint factorises as:

P(X_1, \dots, X_n) = \prod_{i=1}^n P(X_i \mid \text{Parents}(X_i))

Inference answers questions like \( P(\text{Disease} \mid \text{Test} = +) \). Our tool uses enumeration: it sums over all hidden variables to compute the normalizing constant.

Tips for using this page

Keep to 5–6 variables to avoid exponential blowup.
Set every CPT row, or inference will assume 0.5.
Use the evidence toggles to test diagnostic vs causal queries.

1. Nodes

2. Node editor

3. Inference

Bayesian network recap

Tips for using this page