MAT Palestras - Teoria da Probabilidade

Statistical Inference for the Generalized LangevinEquation

Neste seminário, apresentaremos o estimador de máxima verossimilhança (MLE) para o parâmetro drift da equação de Langevin generalizada guiada por um processo L ́evy observado continuamente no tempo.
Felipe Sousa Quintino - UnB em 27/08/2021

Abstract

In this seminar, we will present the maximum likelihood estimator (MLE) for the drift parameter of the generalized Langevin equation driven by a L ́evy process observed continuously in time. Generally, the MLE has a non-explicit form and we present its consistency (Law of Large Numbers), asymptotic normality (Central Limit Theorem) and efficiency (minimum variance). A discretization of the MLE is proposed and estimations from simulated paths were done for the generalized Ornstein-Uhlenbeck process of the fluctuating exponential type.

Local e Data

Tema: Statistical Inference for the Generalized LangevinEquation
Palestrante: Felipe Sousa Quintino - UnB
Data: 27/08/2021