Workshop SDS: Registration
View Submission - SDS2022
A0193
Title: Mixtures, heavy tails, asymmetry and conditional heteroskedasticity in financial returns Authors:  Fariborz Setoudehtazangi - University of Padova (Iran) [presenting]
Massimiliano Caporin - University of Padova (Italy)
Abstract: Statistical modeling and analysis based on finite mixtures of symmetric distributions, especially normal mixtures, have been applied in many applications. In recent years, finite mixtures of asymmetric distributions have been developed as a powerful substitute for normal mixtures in a wide range of applications. The benefit of finite mixture models is to accommodate different characteristics, such as multimodality, skewness, kurtosis and heavy tails. Modeling financial data is considered a complex task, so not only working with financial returns is often involved in such a complex relationship with prior observations, but also the innovations, after fitting an appropriate model, drastically demonstrate skewness, kurtosis, heavy tail and multimodality. Financial returns usually reverberate a structure which can be logically illustrated with conditional heteroskedastic models, such as the Generalized Autoregressive Conditional Heteroskedastic (GARCH) process of Bollerslev. We propose a model adopting a mixture of asymmetric distribution for a financial time series characterized by conditional heteroskedasticity. The stochastic representation of the proposed model enables us to easily implement an EM-type algorithm to estimate the unknown parameters of the model. A comprehensive simulation study and real data sets are then conducted to evaluate the higher performance of the proposed method.