T2V_TF: An adaptive timing encoding mechanism based Transformer with multi-source heterogeneous information fusion for portfolio management: A case of the Chinese A50 stocks

作者：

Highlights：

• A novel model called T2V_TF is proposed to deal with portfolio management.

• Multi-source heterogeneous data, including time-series and text, are considered.

• Time2Vec makes the proposed model more adaptive in positional encoding.

• News sentiment is used in forecast, and is verified its role to investment returns.

• Experiments confirm the validity in terms of T2V_TF and multi-source data fusion.

摘要

•A novel model called T2V_TF is proposed to deal with portfolio management.•Multi-source heterogeneous data, including time-series and text, are considered.•Time2Vec makes the proposed model more adaptive in positional encoding.•News sentiment is used in forecast, and is verified its role to investment returns.•Experiments confirm the validity in terms of T2V_TF and multi-source data fusion.

论文关键词：Portfolio management,Multi-source heterogeneous information,Time2Vec,Transformer,A50 stocks

论文评审过程：Received 31 May 2022, Revised 17 September 2022, Accepted 9 October 2022, Available online 15 October 2022, Version of Record 21 October 2022.

论文官网地址：https://doi.org/10.1016/j.eswa.2022.119020