A novel risk-control model for the online portfolio selection of high-frequency transactions
作者：
Highlights：
• We propose a novel risk-control model for the online portfolio selection.
• The model can conduct bidirectional increment learning.
• The model can remove the system noise at the eigenvector level.
• We propose two new metrics.
• The model can produce excess returns while effectively controlling risk.
摘要
•We propose a novel risk-control model for the online portfolio selection.•The model can conduct bidirectional increment learning.•The model can remove the system noise at the eigenvector level.•We propose two new metrics.•The model can produce excess returns while effectively controlling risk.
论文关键词：Online portfolio selection,Bidirectional-broad learning system,Principal component analysis,Cumulative return,Risk control
论文评审过程：Received 1 May 2021, Revised 26 November 2021, Accepted 4 January 2022, Available online 11 January 2022, Version of Record 31 January 2022.
论文官网地址：https://doi.org/10.1016/j.knosys.2022.108176
