报告时间:2024年5月29日（周三）10:00
报告地点:明德主楼1016
报告主题一:An Efficient Approach to High-Dimensional Portfolio Optimization in Volatile Markets
报告摘要：
With the global financial market experiencing continuous expansion and escalating volatility, the development of efficient strategies for high-dimensional portfolio allocation has become critically important. Previous approaches to high-dimensional portfolio selection have mainly focused on large-cap companies, presenting challenges when confronted with datasets such as the Russell 2000 index. This paper aims to address portfolio optimization challenges within this context, using the 2020-2021 U.S. stock market as a case study. We propose a Dantzig-type portfolio optimization (DPO) model, and present efficient parallel computing algorithms based on asset-splitting. Through empirical analysis on the S&P 500 and Russell 2000 indices, we demonstrate the consistent outperformance of the DPO portfolios over Markowitz mean-variance and Lasso-type mean-variance models, as well as corresponding ETFs, in terms of Sharpe and Sortino ratios. This outperformance is particularly pronounced for the Russell 2000 index. We provide a new effective approach for investors seeking to optimize their portfolios in complex market environments.
作者简介：
杨松山，中国人民大学统计与大数据研究院，任助理教授、博士生导师。研究兴趣包括高维数据分析，模型算法优化，机器学习以及统计模型在金融学、生理学和心理学中的应用。在JASA、JOE、JCGS等国际统计学期刊发表十余篇文章。
报告主题二:A Variable Selection Tree and Its Random Forest
报告摘要：
The Sure Independence Screening (SIS) provides a fast and efficient ranking for the importance of variables for ultra-high dimensional regressions. However, classical SIS cannot eliminate false importance in the ranking, which is exacerbated in nonparametric settings. To address this problem, a novel screening approach is proposed by partitioning the sample into subsets sequentially and creating a tree-like structure of sub-samples called the SIS-tree. SIS-tree is straightforward to implement and can be integrated with various measures of dependence. Additionally, SIS-tree is extended to a forest with improved performance. Through simulations, the proposed methods are demonstrated to have great improvement comparing with existing SIS methods. The selection of a cutoff for the screening is also investigated through theoretical justification and experimental study. As a direct application of the screening, the classification of high-dimensional data is considered, and it is found that the ranking and cutoff can substantially improve the performance of existing classifiers.
报告人简介：
蔡智博，中国人民大学统计学院讲师。博士毕业于新加坡国立大学统计与数据科学系，曾就职字节跳动视频架构部。主要从事高维统计分析、机器学习的研究，主要研究兴趣包括充分降维、变量选择及其在机器学习中的应用，以及生成式人工智能的理论研究与应用。学术论文在JASA、NeurIPS、ICLR等学术期刊会议上发表。
