Editorial
About smartlytics
smartlytics explores decisions under uncertainty, from Bayesian statistics through simulation and optimisation to machine learning and language models.
What you will find here
We explain methods for turning data into good decisions: Bayesian statistics, A/B testing, decision theory, Monte Carlo simulation, optimisation, bandits and Thompson sampling, reinforcement learning and the connecting framework, decision intelligence. Many topics include interactive explainers that calculate directly in your browser.
You can also explore causal inference, bootstrap resampling, regression, clustering, attention and language models with retrieved sources. The learning paths connect methods to practical questions. Change assumptions in the simulations, compare outcomes and explore where each model applies.
Our principles
- Checkable calculations. Formulas are shown in full, and examples can be verified on paper, in a spreadsheet or in the explainer.
- Sources included. Every article identifies the textbooks and original research it draws on. We state where we simplify.
- Illustrative means illustrative. Numerical examples are constructed to demonstrate a principle. They are not measurements, customer results or benchmarks.
- Uncertainty is part of the answer. We report intervals and assumptions, not just point estimates, and describe each method’s limits.
- Kept current. Every article carries the date of its last subject-matter review. We welcome corrections at impressum@nyxai.com.
Publisher
smartlytics is the knowledge resource of NyxAI GmbH in Ingolstadt, Germany. NyxAI develops AI agents and decision systems for demanding business processes. At relevant points, articles explain how NyxAI uses the methods described: discreetly, clearly identified and without requiring you to leave any personal data.
Privacy at a glance
smartlytics uses no cookies, analytics scripts or advertising trackers. Explainers calculate locally. See the privacy policy for details.