Probability and Statistics

The group of Probability and Statistics aggregates researchers from CMUP with the common denominator that they all use tools of probability and statistics to carry out research. 

The research subjects covered include extreme value theory, ergodic theory, signal processing, statistical modelling and inference, machine learning and data mining. Moreover, applications to computer science, biomedicine, engineering, insurance and meteorology are also studied.  

Some members work in Statistics of Extremes (Extreme Value Theory), where the estimation of the tail behaviour is of crucial importance for applications to risk assessment associated with rare but often catastrophic events. Research within this subject involves modelling tail data sets and making statistical inference specially adapted to the extremal context.

Some members of the group belong to the research strand of Dynamical Systems (DS) and develop research in Ergodic Theory. In general words, the contribution from these members for the research accomplished within the group resides in the study of statistical properties of chaotic dynamical systems. This purpose is achieved by using probabilistic tools to investigate systems, which behave erratically and, for that reason, are difficult to forecast, but whose statistical behaviour presents some regularity that contributes for a better understanding of the underlying dynamics.

Other members work in Data Analysis and Signal Processing. Research involves studying classification problems, developing models for data fitting and methodologies to cope with statistical inference and data interpretation.  The research is applied to computer science, medicine and pharmacology. In particular, several case studies include analysing biomedical signals and derived time series.

The group also includes people developing new methodologies for supervised and unsupervised classifications as well as members developing mathematical programming models for real life situations by using methods of machine learning and data-driven optimisation.  

Expectedly, several group members belong to the research strand of Mathematical Models and Applications.