A Risk-Based Trading System Using Algorithmic Trading and Deep Learning Models

Amirmohammad Maleki, Ehsan Hajizadeh () and Ali Fereydooni
Additional contact information
Amirmohammad Maleki: Amirkabir University of Technology, Department of Industrial Engineering and Management Systems
Ehsan Hajizadeh: Amirkabir University of Technology, Department of Industrial Engineering and Management Systems
Ali Fereydooni: Amirkabir University of Technology, Department of Industrial Engineering and Management Systems

A chapter in Data Analytics for Management, Banking and Finance, 2023, pp 135-155 from Springer

Abstract: Abstract The importance of investment is well understood in recent times. Investing and trading in gold is one strategy in this regard, which is facilitated by the commodity market. The depth and speed of this market are increasing every day. However, the ability of individuals to control their emotions, act swiftly, and think rationally is lagging behind the market, resulting in an increased likelihood of traders facing losses. Many large traders, companies, and banks have adopted various statistical, data science, and algorithmic trading models to achieve their goals in financial markets. In this study, we propose an algorithmic trading strategy for the gold asset that serves two objectives: (1) signaling based on predicting risk parameters such as stop loss and take profit, and (2) considering the maximum number of open positions and the capital reduction percentage in each trade. To achieve these objectives, we first employ the long short-term memory (LSTM) neural network to train the best model for predicting gold prices. We use the random forest, permutation, and clustering techniques between the inputs to select the best features. It will be demonstrated that combining feature selection methods and compensating for the lost values significantly improve the model’s results. After predicting the price, we explore and identify the best value for the stop loss, take profit, the maximum number of open positions, and the capital reduction percentage in each trade. In contrast to many previous studies that only consider the return as a factor, we consider different factors to determine the optimal values. Subsequently, we develop an algorithmic trading strategy that incorporates the obtained values and uses the trend prediction method to choose between open positions and the current signal in the opposite direction. Finally, we demonstrate that applying the LSTM neural network to select between the current signal and open positions enhances the trading strategy by predicting the trends.

Keywords: Algorithmic trading; Trading strategy; Risk management; Deep learning (search for similar items in EconPapers)
Date: 2023
References: Add references at CitEc
Citations:

There are no downloads for this item, see the EconPapers FAQ for hints about obtaining it.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:sprchp:978-3-031-36570-6_6

Ordering information: This item can be ordered from
http://www.springer.com/9783031365706

DOI: 10.1007/978-3-031-36570-6_6

Access Statistics for this chapter

More chapters in Springer Books from Springer
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().