High-frequency trading factor mining using genetic programming in China's a-share market

Student thesis: MRes Thesis

Abstract

This research presents an innovative high-frequency factor mining method based on a genetic programming to solve the problem of choosing the trading time in T+0 trading activity in the Chinese stock market. Starting from the underlying logic, this paper discusses why genetic programming are particularly suitable for high-frequency quantitative trading factor mining. Using genetic programming, this research successfully found a factor ideal for high-frequency trading, called the big wave factor. It can evaluate the possibility of significant fluctuations in stock prices in the future. Given that this factor cannot judge the direction of fluctuations, this paper customizes a feasible trading strategy for this factor. Since China’s stock market does not yet support T+0 trading, this trading strategy requires holding a position in the stock in advance, which will affect the calculation of the total return. Therefore, this research proposes a new rate calculation system, especially for intraday trading, which is used to compare with the traditional moving average system strategy. In the experiment, the strategy of this paper has obtained more considerable profits than a traditional strategy.
Date of AwardJul 2022
Original languageEnglish
Awarding Institution
  • University of Nottingham
SupervisorLiang Huang (Supervisor)

Keywords

  • High-frequency trading
  • Quantitative trading
  • Factor mining
  • Genetic programming

Cite this

'