Research Article
Trading Strategies: An Optimal Trading System based on LSTM and Dynamic Programming
@INPROCEEDINGS{10.4108/eai.19-5-2023.2334301, author={Shengyuan Wang}, title={Trading Strategies: An Optimal Trading System based on LSTM and Dynamic Programming}, proceedings={Proceedings of the 2nd International Conference on Bigdata Blockchain and Economy Management, ICBBEM 2023, May 19--21, 2023, Hangzhou, China}, publisher={EAI}, proceedings_a={ICBBEM}, year={2023}, month={7}, keywords={long short-term memory; monte carlo simu- lation; discrete optimization}, doi={10.4108/eai.19-5-2023.2334301} }- Shengyuan Wang
Year: 2023
Trading Strategies: An Optimal Trading System based on LSTM and Dynamic Programming
ICBBEM
EAI
DOI: 10.4108/eai.19-5-2023.2334301
Abstract
This study proposes a series of trading strategies for maximizing total return of Gold and Bitcoin assets over the past five years, while considering transaction commission. The authors preprocess the data by treating the floating prices of Gold and Bitcoin as two stocks, removing missing values, and using time series model LSTM to predict future prices. The LSTM model shows daily price changes in more detail and is selected as the optimal model. Monte Carlo Simulation and Markowitz model are used to find the effective weights of asset combinations, and Dynamic Programming strategy is applied to create an Optimal Action Model for finding the best trading dates. The overall model is found to be sensitive to transaction commission change.
