Knowledge-aware Attentional Neural Network based healthcare big data analytics optimized with Weighted Velocity-Guided Grey Wolf Optimization Algorithm

• Published in: Biomedical signal processing and control Vol. 110; p. 108160
• Main Authors: Vasuki, N., Anand, C., Sukumar, P., Babu, V.Suresh
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.12.2025
• Subjects: Healthcare big data; Improved Bald Eagle Search Optimization Algorithm; Knowledge-aware Attentional Neural Network; Multiparticle Kalman filter; Weighted Velocity-Guided Grey Wolf Optimization Algorithm; Multiparticle Kalman filter; Knowledge-aware Attentional Neural Network; Healthcare big data; Weighted Velocity-Guided Grey Wolf Optimization Algorithm; Improved Bald Eagle Search Optimization Algorithm
• ISSN: 1746-8094
• DOI: 10.1016/j.bspc.2025.108160

A significant increase in data volumes, along with the attractive opportunities and potential arising from data analysis contributes to the idea of Big Data. The existing healthcare big data analytics methods face challenges in handling high-dimensional data, slow convergence and suboptimal feature selection. In this paper, a Knowledge-aware Attentional Neural Network based Healthcare Big Data Analytics optimized with Weighted Velocity-Guided Grey Wolf Optimization Algorithm (KANN-HBA-WVGGWOA) is proposed. Here, the input data are taken from PIMA Indians Diabetes dataset. Then the input data is pre-processed by utilizing Multiparticle Kalman filter (MKF) to calculate every data object value primarily. The feature selection utilizing Improved Bald Eagle Search Optimization Algorithm (IBESOA) to select the optimal features from the dataset. The selected features are given into Knowledge-aware Attentional Neural Network (KANN) to classify the data as diabetes and no diabetes. Finally, Weighted Velocity-Guided Grey Wolf Optimization Algorithm (WVGGWOA) is proposed to optimize the KANN classifier that precisely classifies the diabetes disease. The KANN-HBA-WVGGWOA method is implemented in Python. The proposed KANN-HBA-WVGGWOA method attains 1.28%, 2.22%, and 2.27% higher accuracy; 12.56%, 18.68%, and 19.49% less computational time compared to the existing models: Role of big data analytics for revolutionizing diabetes management including health care decision-making (BDA-LR-RDMH), Map reduce dependent big data framework utilizing associative kruskal poly kernel classifier for diabetic disorder prediction (BDF-MRPK-DDP) and the Implementation of ML approaches with big data along IoT to generate effectual prediction for health informatics (BD-KNN-PHI) respectively.