报告人：黄磊 副教授 (西南交通大学)

时间：2026年4月25日 10:30-

地址：数统学院LD718

摘要：We propose RM-SPAGL, a high-dimensional portfolio selection method that combines the equivalent regression formulation of mean variance optimization with Sparse Group Lasso regularization and factor based covariance structure. The procedure incorporates industry grouping, allows simultaneous group and within group sparsity, and selects the tuning parameter by risk-constrained cross-validation to target a specified portfolio risk level. Under standard regularity conditions, we establish convergence of the resulting portfolio return to its theoretical target at a rate determined by both element-wise and group level sparsity. Simulation results show that RM-SPAGL achieves near optimal Sharpe ratios, accurate risk control, and negligible white noise selection. Out-of-sample analyses using S&P 500 constituents and Chinese A-Share Market further show that the proposed method delivers competitive risk adjusted performance with substantially lower turnover than benchmark procedures. These results indicate that exploiting group structure can improve the stability, interpretability, and practical performance of high-dimensional portfolio selection.

邀请人：夏小超

欢迎广大师生积极参与！