報(bào)告人: 張新雨 研究員

講座日期：2019-12-11

講座時(shí)間：10:00

報(bào)告地點(diǎn)：長(zhǎng)安校區(qū) 數(shù)學(xué)與信息科學(xué)學(xué)院學(xué)術(shù)交流廳

主辦單位：數(shù)學(xué)與信息科學(xué)學(xué)院

講座人簡(jiǎn)介：

張新雨，中科院系統(tǒng)所/預(yù)測(cè)中心研究員，Texas A&M大學(xué)博士后、Penn State 大學(xué)Research Fellow。主要研究方向?yàn)槟Ｐ推骄?、模型選擇、組合預(yù)測(cè)等。先后主持杰青、優(yōu)青等4項(xiàng)國(guó)家自然科學(xué)基金，目前擔(dān)任《JSSC》、《SADM》、《系統(tǒng)科學(xué)與數(shù)學(xué)》、《應(yīng)用概率統(tǒng)計(jì)》編委和《Econometrics》客座主編。

講座簡(jiǎn)介：

In this paper, we develop a model averaging method to estimate the high-dimensional covariance matrix, where the candidate models are constructed by different orders of the polynomial functions. We propose a Mallows-type model averaging criterion and select the weights by minimizing this criterion, which is an unbiased estimator of the expected in-sample squared error plus a constant. Then, we prove the asymptotic optimality of the resulting model average covariance (MAC) estimators. Furthermore, numerical simulations and a case study on Chinese airport network structure data are conducted to demonstrate the usefulness of the proposed approaches.