Read the full-text of this article for free

Title: A particle swarm optimisation-based deep belief network for traditional Chinese medicine data processing strategies

Authors: Wenqiao Ding; Chongli Xu; Ruiqi Zhang

Addresses: School of Biological and Food Engineering, Jilin University of Chemical Technology, Jilin, 132022, China ' School of Medical Technology, Chongqing Medical and Pharmaceutical College, Chongqing, 401331, China ' Faculty of Business, Economic and Accountancy, University Malaysia Sabah, Jalan UMS, 88400, Malaysia; School of Biological and Food Engineering, Jilin University of Chemical Technology, Jilin, 132022, China

Abstract: To address the challenges of high dimensionality, nonlinearity, and small sample size in traditional Chinese medicine data, this study proposes a novel data processing strategy using a particle swarm optimised deep belief network. The method automates deep belief network hyperparameter tuning via particle swarm optimised to enable end-to-end feature learning and classification. On the traditional Chinese medicine systems pharmacology Danshen blood-activation dataset, particle swarm optimised deep belief network achieved an accuracy of 87.8% and an F1-score of 86.8%, surpassing both conventional models, (e.g., XGBoost at 85.2%) and unoptimised deep belief network (83.1%). It also attained 98.8% accuracy on the University of California Irvine Wine dataset, demonstrating strong generalisation. This work offers an automated, high-precision computational tool for traditional Chinese medicine data analysis, significantly enhancing model performance and interpretability.

Keywords: particle swarm optimisation; deep belief networks; DBNs; traditional Chinese medicine data; hyperparameter optimisation; feature learning.

DOI: 10.1504/IJICT.2025.151059

International Journal of Information and Communication Technology, 2025 Vol.26 No.49, pp.36 - 56

Received: 25 Sep 2025
Accepted: 26 Oct 2025
Published online: 12 Jan 2026 *