BAYESLAB · LEARNING RESOURCES
BLOGS
Make probability click. Practical walkthroughs, worked examples, and modelling ideas for students who learn by doing.
Start small. Build your intuition.
New to the subject? Begin with your first Bayesian network, then work through conditional tables, evidence, and decisions. Each walkthrough uses explicit teaching assumptions so you can reproduce the result and experiment.
Walkthroughs & ideas
Bayes’ theorem, without the mystery
Work through a faulty-parcel detector example to understand priors, likelihoods, false positives, and posterior probability.
6 min read · Worked exampleRead the walkthroughConditional probability tables: get every row right
Build a two-parent delivery model, fill its complete CPT, and calculate a marginal probability by weighting the parent combinations.
6 min read · Worked exampleRead the walkthroughExplaining away: when one cause makes another less likely
Use a two-cause alarm model to see how observing a shared effect connects otherwise independent variables. Includes a full CPT and posterior calculations.
7 min read · Worked exampleRead the walkthroughConditional independence: when more evidence stops helping
Build a weather–traffic–lateness chain and verify how observing the middle variable blocks information from the upstream variable.
6 min read · Worked exampleRead the walkthroughFrom probabilities to decisions: expected utility
Compare two fictional event plans with an influence diagram, calculate expected utility, and find the probability where the preferred choice changes.
6 min read · Worked exampleRead the walkthroughYour model gives a surprising answer. What now?
A practical Bayesian network debugging walkthrough: check state labels, CPT totals, evidence, direction, and a hand-calculated reference result.
6 min read · Worked exampleRead the walkthrough