Skip to content

This research aims to bridge this gap by investigating the feasibility and effectiveness of integrating deep learning methods into an algorithmic trading system specifically tailored for cryptocurrency trading

Notifications You must be signed in to change notification settings

tomaspzpeborges/TradeAI

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 

Repository files navigation

Please find complete work in the original repository

TradeAI

Integrating deep learning methods into algorithmic trading systems is advancing the financial industry, enabling sophisticated analysis and decision-making capabilities previously limited to institutional investors, making them accessible to retail investors. However, the application of deep learning in medium- frequency trading, particularly in the cryptocurrency market, remains largely unexplored. This research aims to bridge this gap by investigating the feasibility and effectiveness of integrating deep learning methods into an algorithmic trading system specifically tailored for cryptocurrency trading. This study presents a comprehensive full-stack algorithmic trading system that integrates deep learning methods for trading strategies. Specifically, we delve into utilising novel transformer architectures, including Informer, Pyraformer, and an enhanced original transformer, into predicting cryptocurrency prices. Our research findings showcase these transformer models’ superior performance compared to traditional ARIMA models, particularly when operating on larger datasets. Notably, the Pyraformer model exhibits exceptional predictive accuracy while maintaining efficient training and inference times. Moreover, we seamlessly integrate these predictive signals into the environment definition of a Deep Reinforcement Learning (DRL) model, enabling effective order generation and decision-making. The findings contribute to understanding transformer models’ effectiveness in medium-frequency cryptocurrency price prediction and provide a promising architecture for future research and development in this evolving field.

Context

This work was developed as the final year thesis for MEng Computer Science at the University of Leeds. Tomás ZB achieved a project grade of 87 and the work was consired fit for publishing.

About

This research aims to bridge this gap by investigating the feasibility and effectiveness of integrating deep learning methods into an algorithmic trading system specifically tailored for cryptocurrency trading

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published