The Bayesian learning module introduces how learning and decision-making processes under uncertainty can be modelled through a Bayesian approach. Specifically, it will first illustrate how experts' knowledge can be implemented into probability distributions, within a robust Bayesian approach. Second, it will go through the learning process of Bayesian networks, which allow to express very complex problems through a set of related smaller ones, based on dependency and probabilistic relations. Third, it will illustrate how machine learning of large amount of data can be made efficient by means of Bayesian classifiers. Finally,it will address how Bayesian learning can be used in the context of adversarial risk analysis and cyber security.
Prerequisiti
Basic Knowledge of Statistics and/or Probability. Basic Knowledge of Coding
Metodi didattici
Theory and practical examples.
Verifica Apprendimento
Oral Discussion
Testi
- Banks, D., Rios, J., and Rios Insua, D. (2015). Adversarial Risk Analysis (Vol. 343). CRC Press. - O'Hagan, A., Buck, C.E., Daneshkhah, A., Eiser, J.R., Garthwaite, P.H., Jenkinson, D.J., Oakley, J.E, and Rakow, T. (2006). Uncertain Judgements: Eliciting Experts' Probabilities. Wiley. - Rios Insua, D. and Ruggeri, F. Eds. (2000), Robust Bayesian Analysis, Lecture Notes in Statistics, vol. 152, Springer. - Materials: data, code and papers supplied in class for the financial learning part.
Contenuti
Introduction to Bayesian Statistics and Decision Analysis - Elicitation of experts' opinions and Bayesian Robustness - Bayesian networks - Bayesian classifiers - Adversarial Risk Analysis and classification - Adversarial classification
