Auto uning of price prediction models for high-frequency trading via reinforcement learning
作者：
Highlights：
• Construct a light model library, train different light model for each market distribution.
• Devise a novel reward function based on inverse reinforcement learning, and estimate the profits of each order .
• Devise an algorithm for dynamically selecting model based on reinforcement learning.
• Promising performance on real-world high-frequency trading in China Commodity Future market.
摘要
•Construct a light model library, train different light model for each market distribution.•Devise a novel reward function based on inverse reinforcement learning, and estimate the profits of each order .•Devise an algorithm for dynamically selecting model based on reinforcement learning.•Promising performance on real-world high-frequency trading in China Commodity Future market.
论文关键词：High-frequency trading,Inverse reinforcement learning,Parameter optimization,Multi-armed bandit
论文评审过程：Received 30 June 2020, Revised 14 December 2021, Accepted 16 January 2022, Available online 18 January 2022, Version of Record 24 January 2022.
论文官网地址：https://doi.org/10.1016/j.patcog.2022.108543
