• <label id="xsx3j"><ruby id="xsx3j"></ruby></label>
  • <mark id="xsx3j"></mark>

      <mark id="xsx3j"></mark>
    1. 新聞中心NEWS


      講座:Structural-Factor Modeling of Big Dependent Data

      發布者:人力資源辦公室    發布時間:2020-12-22

      題 目:Structural-Factor Modeling of Big Dependent Data

      演講人:高照省  助理教授     Lehigh University

      主持人:李成璋  助理教授  上海交通大學安泰經濟與管理學院

      時 間:2020年12月24日(周四) 10:40-12:10



      In this talk, we will introduce a new approach to modeling high-dimensional time series by treating a p-dimensional time series as a nonsingular linear transformation of certain common factors and idiosyncratic components. Unlike the approximate factor models, we assume that the factors capture all the non-trivial dynamics of the data, but the cross-sectional dependence may be explained by both the factors and the idiosyncratic components. Under the proposed model, (a) the factor process is dynamically dependent and the idiosyncratic component is a white noise process, and (b) the largest eigenvalues of the covariance matrix of the idiosyncratic components may diverge to infinity as the dimension p increases. We propose a white noise testing procedure for high-dimensional time series to determine the number of white noise components and, hence, the number of common factors, and introduce a projected Principal Component Analysis (PCA) to eliminate the diverging effect of the idiosyncratic noises. Asymptotic properties of the proposed method are established for both fixed p and diverging p as the sample size n increases to infinity. We use both simulated data and real examples to assess the performance of the proposed method. We also compare our method with two commonly used methods in the literature concerning the forecastability of the extracted factors and find that the proposed approach not only provides interpretable results, but also performs well in out-of-sample forecasting. The results suggest that the dynamically dependent factors extracted by our method fare better in predicting the U.S. Consumer Price Indexes. 


      Zhaoxing Gao is a tenure-track Assistant Professor in Department of Mathematics at Lehigh University in Pennsylvania, USA. He is also a faculty member of Institute for Data, Intelligent Systems & Computation at Lehigh. Before joining Lehigh, he got his Ph.D. from the Hong Kong University of Science and Technology in 2016,  and worked as a Postdoc research officer at London School of Economics and Political Science from 2016 to 2017, and a Postdoc research Professional at Booth School of Business, University of Chicago from 2017 to 2019. His research interest is focused on Data Science, Machine Learning and Big Data in Finance, Business Statistics and Fintech; Factor-based Investment and Forecasting; Risk Management and Credit Ratings.