Skip to content
All library documents

Bayesian Optimization for Adaptive Nonstationary Trading Strategies

Article arXiv papers · Author: Bingde Liu et al.
Read the original

Summary

This paper frames automated trading under changing market conditions as a nonstationary continuum-armed bandit problem. It proposes PRBO, which combines Bayesian optimization with a bandit-over-bandit approach to adjust strategy parameters as conditions change. The method is evaluated in the Bristol Stock Exchange simulation, where automated agents with heterogeneous strategies populate the market. Its benchmark is PRSH, a strategy that adapts parameters using stochastic hill-climbing.

The reported simulation results show PRBO earning significantly more profit than PRSH while requiring fewer hyperparameters to tune. This suggests a possible advantage for Bayesian optimization in the tested adaptive-trading setup. However, the supplied description gives no absolute returns, uncertainty estimates, or details about the market scenarios and parameter settings. The evidence is limited to the stated simulated comparison, so it does not establish that PRBO will outperform in live markets or under other benchmarks and conditions.

Key ideas

  • The paper models adapting trading parameters to changing conditions as a nonstationary continuum-armed bandit problem.
  • PRBO uses Bayesian optimization within a bandit-over-bandit framework to adjust strategy parameters.
  • The evaluation uses the Bristol Stock Exchange simulation with heterogeneous automated agents.
  • PRBO is compared with PRSH, which adapts parameters through stochastic hill-climbing.
  • The reported simulation gives PRBO significantly higher profit with fewer hyperparameters to tune.
  • The excerpt does not establish performance in live markets or across other benchmarks.

Tags

Machine learningBacktestingEquitiesStatistics
Full text
# Nonstationary Continuum-Armed Bandit Strategies for Automated Trading in a Simulated Financial Market


# Nonstationary Continuum-Armed Bandit Strategies for Automated Trading in a Simulated Financial Market









We approach the problem of designing an automated trading strategy that can consistently profit by adapting to changing market conditions. This challenge can be framed as a Nonstationary Continuum-Armed Bandit (NCAB) problem. To solve the NCAB problem, we propose PRBO, a novel trading algorithm that uses Bayesian optimization and a ``bandit-over-bandit'' framework to dynamically adjust strategy parameters in response to market conditions. We use Bristol Stock Exchange (BSE) to simulate financial markets containing heterogeneous populations of automated trading agents and compare PRBO with PRSH, a reference trading strategy that adapts strategy parameters through stochastic hill-climbing. Results show that PRBO generates significantly more profit than PRSH, despite having fewer hyperparameters to tune. The code for PRBO and performing experiments is available online open-source (https://github.com/HarmoniaLeo/PRZI-Bayesian-Optimisation).

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.