Joint propagation of ontological and epistemic uncertainty across risk assessment and fuzzy time series models

Abstract

This paper discusses hybrid probabilistic and fuzzy set approaches to propagating randomness and imprecision in risk assessment and fuzzy time series models. Stochastic and Computational Intelligence methods, such as Probability bounds analysis, Fuzzy α-levels analysis, Fuzzy random vectors, Wavelets decomposition and Wavelets Networks are combined to capture different kinds of uncertainty. Their most appropriate applications are probabilistic risk assessments carried out in terms of probability distributions with imprecise parameters and stochastic processes modeled in terms of fuzzy time series.