Talk 2
One Day Meeting on Advanced Statistics Topics
CIDMA- University of Aveiro
Speaker: Maria da Conceição Costa - CIDMA & DMat, Universidade de Aveiro
Title of the talk: Non linear time series models and the application of optimum alarm systems
Abstract: In this talk, some classes of models usually applied in the analysis of time series of real values and integer values will be discussed, with a focus on their application to the analysis of financial series. Among the models with conditional heteroscedasticity, special attention is given to the Fractionally Integrated Asymmetric Power ARCH, FIAPARCH(p,d,q), and an optimal alarm system is implemented, considering both the classical and Bayesian methodologies.
Considering the particular characteristics of financial time series and the properties of the APARCH(p,q) model, the proposed homologue for the modeling of time series of counts is presented: the INteger-valued Asymmetric Power ARCH, INAPARCH(p,q). The probabilistic properties of the INAPARCH(1,1) model are presented, the conditional maximum likelihood (CML) estimation method is applied, and the asymptotic properties of the CML estimator are studied. The implementation of an optimal alarm system to the INAPARCH(1,1) model is also described and an application to real data series is presented.
Maria da Conceição Costa completed her PhD in Mathematics in January 2014 by Universidade de Aveiro. She is currently Assistant Professor in Universidade de Aveiro, Departamento de Matemática and researcher at the CIDMA (Center for Research and Development in Mathematics and Applications), at the same university. Her main areas of research are time series analysis with focus on surveillance systems, maximum-entropy based inference and applications to economics and health sciences.
ORCID: 0000-0002-4776-6375