Algorithmic Trading Strategies and Simulations for Multi-asset Classes

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering


PhD Thesis Defence


Title: "Algorithmic Trading Strategies and Simulations for Multi-asset Classes"

By

Mr. Feng HAN


Abstract

In modern limit order book(LOB)-driven financial markets, we contribute 
algorithmic trading strategies for multi-asset classes, including traditional 
assets, equities, futures contracts, and emerging assets like cryptocurrencies. 
With the capabilities of a computer system, algorithmic trading can be applied 
to short-term, even minute intervals.

The first contribution is algorithmic trading strategies for Hong Kong stocks. 
We generate trading signals at the minute intervals, implement SAR & MACD 
modules and backtest 391 stocks. The second contribution is statistical 
arbitrage across different futures exchanges. We explore precious metal futures 
contracts listed on Shanghai Futures Exchange, Chicago Mercantile Exchange, and 
Japan Exchange Group. Our results show the correlation observations for futures 
contracts in 2020. Apart from traditional asset classes, the third contribution 
is cryptocurrency spots and perpetual futures trading with technical indicators 
including multi-strategies, e.g., shorting. We backtest 2019-2021 multiple 
periods. To understand the underlying distribution of each asset, we create a 
simulator to generate orders based on Wasserstein GAN as the fourth 
contribution. We compare the original data with the generated real and fake 
data using Kolmogorov-Smirnov tests, MSE, MAE to quantify our model 
performance.

Multi-asset classes investment boosts overall portfolio performance with risks 
diversification across multiple classes. Here we compare three assets: 
equities, futures, and cryptocurrency from volatility and liquidity. The 
combination of adopting minute intervals observations with a systematic 
computation across equities, commodity futures, and emerging cryptos can be an 
exciting topic. This thesis lays out a structure for algorithmic trading 
strategies and simulations.


Date:			Wednesday, 23 February 2022

Time:			3:30pm - 5:30pm

Zoom Meeting: 		https://hkust.zoom.us/j/2871974170

Chairperson:		Prof. Daniel PALOMAR (ECE)

Committee Members:	Prof. Jiheng ZHANG (Supervisor, IEDA)
 			Prof. Xiaojuan MA (Supervisor)
 			Prof. Qifeng CHEN
 			Prof. Qiong LUO
 			Prof. Kani CHEN (MATH)
 			Prof. Yi FANG (Jilin University)


**** ALL are Welcome ****