Conditional density estimation using fuzzy GARCH models

Abstract

Time series data exhibits complex behavior including non-linearity and path-dependency. This paper proposes a flexible fuzzy GARCH model that can capture different properties of data, such as skewness, fat tails and multimodality in one single model. Furthermore, additional information and simple understanding of the underlying process can be provided by the linguistic interpretation of the proposed model. The model performance is illustrated using two simulated data examples. © 2013 Springer-Verlag.