Article
Computer Science, Artificial Intelligence
Mahdi Massahi, Masoud Mahootchi
Summary: This paper proposes a novel intraday algorithmic trading system for volatile commodity futures markets based on a Deep Q-network algorithm. Experimental results demonstrate that the system outperforms benchmarks in terms of return, risk, and risk-adjusted return.
EXPERT SYSTEMS WITH APPLICATIONS
(2024)
Article
Mathematics, Interdisciplinary Applications
Cohen Gil
Summary: In this research, AI trading systems for intraday trading of five precious metals were designed and optimized. The systems outperformed the B&H returns for Gold, Silver, Platinum and Palladium trades, delivering significant excess returns. Specific configurations of the systems were found to provide better trading returns compared to the buy and hold strategy.
CHAOS SOLITONS & FRACTALS
(2022)
Article
Computer Science, Artificial Intelligence
Thibaut Theate, Damien Ernst
Summary: This research paper introduces an innovative approach based on deep reinforcement learning to address algorithmic trading in the stock market, producing optimized trading strategies and promising results in performance evaluation.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Computer Science, Artificial Intelligence
Zhen Huang, Ning Li, Wenliang Mei, Wenyong Gong
Summary: This paper proposes a new algorithmic trading method called CR-DQN, which combines deep Q-learning with popular trading rules such as moving average and trading range break-out. By generating combinational rule vectors using a reward driven combination weight updating scheme and designing a piecewise reward function, this method exhibits the best performance on Chinese and non-Chinese stock markets.
APPLIED SOFT COMPUTING
(2023)
Article
Economics
Yingjian Pu, Baochen Yang
Summary: This study proposes a new trading strategy and analyzes its profitability in both the Chinese and US commodity futures markets. The results show that the historical basis strategy performs better than the momentum strategy but worse than the carry strategy. Furthermore, incorporating energy futures into trading strategies and portfolios significantly increases profitability but also raises risks.
Article
Computer Science, Cybernetics
Yuze Li, Shouyang Wang, Yunjie Wei, Qing Zhu
Summary: This study proposes a new hybrid forecasting approach for gold price prediction, capable of extracting internal factors and patterns, detecting changes in market conditions, and accurately predicting price fluctuations. Experimental results demonstrate significant improvements in prediction performance compared to benchmarks, with consistent positive returns in trading strategy generation and testing over an 11-year out-of-sample period. The approach also shows promising results when applied to the spot gold market, offering practical guidance for managing investment risk and implementing hedging strategies in the gold commodity market.
IEEE TRANSACTIONS ON COMPUTATIONAL SOCIAL SYSTEMS
(2021)
Article
Computer Science, Artificial Intelligence
Ali Shavandi, Majid Khedmati
Summary: This study proposes an algorithmic trading framework based on machine learning that utilizes the collective intelligence of multiple agents for robust trading in financial markets. Through experimental evaluation on a historical dataset, the framework outperforms independent agents and benchmark trading strategies across multiple trading timeframes.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Huang Ke, Zhang Zuominyang, Li Qiumei, Luo Yin
Summary: This paper proposes an EEMD-Hurst-LSTM prediction method based on the ensemble learning framework for the prediction of typical commodities in China's commodity futures market. The method utilizes EEMD and the adaptive fractal Hurst index to incorporate new features into the LSTM model, improving its correlation detection with the external market. The results show that the EEMD-Hurst-LSTM method outperforms other models in predictive performance and provides a superior trading strategy with better returns and risk control.
Article
Computer Science, Information Systems
Boyi Jin
Summary: This paper proposes an innovative algorithm to solve the optimal portfolio problem in stock market trading activities. The proposed model outperforms benchmark strategies by utilizing a mean-VaR portfolio optimization model based on the actor-critic architecture, implementing short selling through a linear transformation function, and using a multi-process method to accelerate deep reinforcement learning training.
Article
Computer Science, Artificial Intelligence
Leonardo Kanashiro Felizardo, Francisco Caio Lima Paiva, Catharine de Vita Graves, Elia Yathie Matsumoto, Anna Helena Reali Costa, Emilio Del-Moral-Hernandez, Paolo Brandimarte
Summary: This paper investigates the relationship between machine learning and financial markets, focusing on the application of supervised learning and reinforcement learning approaches in active asset trading. The results suggest that supervised learning can outperform reinforcement learning in trading.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Computer Science, Information Systems
Deog-Yeong Park, Ki-Hoon Lee
Summary: Algorithmic trading allows investors to avoid emotional decisions and make profits with modern technology. Two main challenges in algorithmic trading are extracting robust features and learning profitable trading policies.
Article
Business, Finance
Xingguo Luo, Yuting Lin, Xiaoli Yu, Feng He
Summary: This paper finds that trading activities in the futures option markets have incremental predictive power for commodity futures returns, particularly for both commercial and noncommercial traders. The nonmomentum of position changes in the options market has a significant impact on noncommercial trading compared to idiosyncratic risk and hedging pressure.
JOURNAL OF FUTURES MARKETS
(2021)
Article
Computer Science, Information Systems
Hsien-Ming Chou, Chihli Hung
Summary: This study presents multiple investment strategies for day traders based on visual trend bands on short-term stock index futures, evaluating their performance using machine learning algorithms and deriving strategies for better predictions, with an accuracy rate of 82%.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Energy & Fuels
Zhen Huang, Wenyong Gong, Junwei Duan
Summary: In this paper, a novel deep reinforcement learning algorithm (TBDQN) is proposed to automatically generate consistently profitable and robust trading signals in crude oil and natural gas futures markets.
Article
Environmental Studies
Sangram Keshari Jena, Amine Lahiani, Aviral Kumar Tiwari, David Roubaud
Summary: This study uses a novel approach to analyze the asymmetric relationship between trading activity and commodity futures prices, revealing that the price effect varies across different market conditions. The findings provide insights for portfolio managers, traders, hedgers, and regulators to make effective decisions and policies based on market conditions.