Year of Award
2022
Document Type
Thesis
Degree Type
Master of Science (MS)
Degree Name
Mathematics
Other Degree Name/Area of Focus
Data Science
Department or School/College
Mathematical Sciences
Committee Chair
Javier Perez-Alvaro
Committee Co-chair
Emily Stone
Keywords
Deep learning, Trading, Forex
Publisher
University of Montana
Subject Categories
Data Science
Abstract
Buying and selling Stocks, Foreign Currencies (FOREX), Commodities, and Cryptocurrencies have been a source of wealth generation, and more often, wealth loss for many brave enough to enter the financial markets. In this paper, the author builds on the work of Williams, J. 2022 and develops an agent-based method to solve this wealth generation problem with the use of neural networks. The author points out some assumptions made by Williams, J. 2022 that were sound in theory, but made the implementation of the algorithm presented in their paper diverge from the theory. The author proposes a fundamentally different algorithmic method, called Double Deep Q learning, to trade in financial markets. This method, in the category of reinforcement learning, was popularized when teaching computers how to play video games such as Atari Space Invaders and Pong. The author creates an environment wherein the markets are treated like a video game and an agent learns to find the optimal path through the market to maximize its rewards. The author presents a final algorithm using two (2) interlinked LSTM Recurrent Neural Networks to form a Double Deep Recurrent Q network (DDRQN) algorithm and uses that algorithm to make profitable decisions and secure overall profits in foreign currency exchange markets.
Recommended Citation
Coombs, Sebastian, "TRADING FINANCIAL INSTRUMENTS LIKE A VIDEO GAME: SEARCHING FOR PROFIT USING DEEP REINFORCEMENT LEARNING." (2022). Graduate Student Theses, Dissertations, & Professional Papers. 12051.
https://scholarworks.umt.edu/etd/12051
Included in
© Copyright 2022 Sebastian Coombs