Machine learning with Belief Rule-Based Expert Systems to predict stock price movements
作者:
Highlights:
• Used Belief Rule-Based Expert System to predict the stock price movement.
• Each expert system has a knowledge-base that is created from the training samples.
• Assigns a degree of belief in accordance with the probability of price movement.
• Performance increases when parameters are optimized using the fmincon function.
• Produces promising performance compared to other deep learning approaches.
摘要
•Used Belief Rule-Based Expert System to predict the stock price movement.•Each expert system has a knowledge-base that is created from the training samples.•Assigns a degree of belief in accordance with the probability of price movement.•Performance increases when parameters are optimized using the fmincon function.•Produces promising performance compared to other deep learning approaches.
论文关键词:Stock prediction,Bollinger band,Belief rule,Expert system,Machine learning,Time series analysis
论文评审过程:Received 25 October 2021, Revised 23 April 2022, Accepted 29 May 2022, Available online 4 June 2022, Version of Record 27 June 2022.
论文官网地址:https://doi.org/10.1016/j.eswa.2022.117706