主题：Portfolio Selection with Noisy Return Data
主讲人：舒连杰 教授（澳门大学）
主持人：王国长 教授（暨南大学）
会议工具：腾讯会议（ID：763-994-732）
会议时间：2022年9月19日（周一）10：00-11：30
摘要
In modern portfolio selection, financial investors in reality, especially institutional ones, routinely work with a large number of assets. Moreover, asset return data are often noisy. To meet these challenges, this paper develops a new latent factor model for high-dimensional portfolio selection under noisy asset returns. The new model employs a cellwise method for robust estimation of latent factors and a diagonally-dominant covariance structure to account for cross-sectional dependence in the idiosyncratic component. It has the advantage to be computationally efficient and nearly tuning-free. Comparing favorably with state-of-the-art competitors, a portfolio formed on the new model is able to achieve out-of-sample risk reduction and improve certainty equivalent returns after transaction costs across a wide range of real data sets.
主讲人简介
舒连杰，澳门大学工商管理学院教授，博士课程主任。他1998年获得西安交通大学机械工程和自动化学士学位后，获国家教委推荐赴香港科技大学工业工程及工程管理系直博。舒教授以第一作者或通讯作者在金融工程、工业工程、质量工程及统计等领域国际期刊（如Journal of Financial Quantitative Analysis,Quantitative Finance,IISE Transaction,Journal of Quality Technology,Naval Research Logistics,Statistics in Medicine等）发表60多篇高水平研究论文。其主要研究兴趣为：资产组合优化、高维统计和监控，质量控制和管理及统计计算。舒教授目前担任Journal of Statistical Computation and Simulation副主编。
