High Frequency Quant Trading Algorithm

This project studies on using machine learning to predict high frequency stock movement, and profit from the prediction.

The function to be approximated by XGBoost model is a function that takes order book data and projects the VWAP change in the future 30 seconds.

After a successful approximation of the etf price movement, this project focuses on the challenge of beating the market by high frequency trading.

The back test shows a steady 5% return on a scale of 60 million fund from period 2021-02-01 to 2021-04-01.

Data is from an ETF in Shanghai Stock Exchange.

Jiayang Nie
Jiayang Nie
Machine Learning Engineer Intern