SEMINARIO 09-02-2024: “Empirical Likelihood Methods for Matrix-Valued Time Series with Long Memory"
Venerdì 9 febbraio 2024, ore 12:00
Il Dipartimento di Economia, Statistica e Finanza è lieto di invitarvi al seminario sul tema:
"Empirical Likelihood Methods for Matrix-Valued Time Series with Long Memory"
Relatore: Prof. Yan LIU, Waseda University, Tokyo, Japan
Data e ora: Venerdì 9 febbraio 2024, ore 12:00
Luogo: Aula Seminari DESF, Cubo 0/C, Piano terra.
Abstract: The matrix-valued time series data are often observed in many fields such as economics. To analyze such data, some models, for example, matrix autoregressive model and matrix factor models have been considered. However, such models are considered only in short memory settings and statistical inferences are mainly studied in time domain. In this paper, on the other hand, estimation and testing problem in frequency domain for matrix-valued time series with long term dependence structure are considered. Using the vectorization and considering the frequency domain of the time series, we develop the empirical likelihood method to construct the estimator for the unknown parameter and model diagnostics.
(Joint work with K. Fujimori)
Si allega locandina dell'evento.
tags
Next events
- [[ event.event.publication_data.title ]]
[[ new Date(event.event.date_start).toLocaleString("en", {day: '2-digit', month: '2-digit', year: 'numeric', hour: '2-digit', minute:'2-digit'}) ]][[ event.event.location ]]