PRIN 2022

The general goal of the project is to design interpretable decision models that essentially rest upon the concepts of “ambiguous beliefs”, “vagueness” and “conditioning”. By ambiguous beliefs we mean the simultaneous treatment of classes of probability measures (or their envelopes), by vagueness we intend the encoding of human language expressions, and by conditioning we refer to the capability of considering different information contexts and of updating. The project aims at giving a multidisciplinary approach for integrating decision models, learning and numerical methods. This will provide new insights into economic applications and identify new targets and strategies to justify anomalies in the markets and inform decision making for the benefit of social community.

• T1: Study of conditioning for non-additive uncertainty measures and dynamic hybrid models

• T2: Development of normative decision models under ambiguous beliefs

• T3: Design of data fusion techniques in the framework of imprecise probabilities

• T4: Extension of stochastic processes theory under ambiguous beliefs

• T5: Study of portfolio selection and pricing under partial knowledge

• T6: Design of methods for micro-expression identification for fraud detection