AHBO-DMN: Autoregressive Honey Badger Optimization Deep Maxout Network for Credibility Assessment of Web-Based Health Information
Piyush BaglaKuldeep Kumar1 Dr. B R Ambedkar National Institute of Technology,Jalandhar,144027,India2 National Institute of Technology,Kurukshetra,136119,IndiaPiyush Bagla is pursuing his PhD at Dr. B.R. Ambedkar National Institute of Technology,Jalandhar,Punjab,India,in the Computer Science and Engineering Dept. His research area includes ML and NLP. Before joining NITJ,he worked as an assistant professor at Graphic Era Hill University,Dehradun,Uttarakhand. He has several publications in reputed international journals and conferences. Corresponding author. Email: ldeep Kumar received his PhD from the School of Computing at the National University of Singapore,Singapore 2016. He is an assistant professor in the Department of Computer Engineering at the National Institute of Technology,Kurukshetra. Before that,he worked at the Birla Institute of Technology and Science Pilani and the National Institute of Technology Jalandhar for over six years. His research interests include software engineering,machine learning,and data analytics. He has published over 50 papers in peer-reviewed journals,international conferences,and several book chapters. Email: kuldeepkumar@nitkkr.ac.in
DOI: https://doi.org/10.1080/03772063.2023.2301653
IF: 1.8768
2024-01-09
IETE Journal of Research
Abstract:The cure and treatment suggested by non-medical experts on the Internet can lead to more damage than expected since spreading such misinformation may adversely affect public health. Different interventions have been modeled to help users better evaluate web-based health information. However, retrieving information credibility still results in a complex task. This paper aims to design a novel credibility assessment method using the proposed Autoregressive Honey Badger Optimization-based Deep Maxout Network (AHBO-based DMN) in web-based healthcare data. The AHBO is devised by inheriting Conditional Autoregressive Value At Risk (CAViaR) features with Honey Badger Algorithm (HBA). Finally, DMN, which is trained by the developed AHBO algorithm, accomplishes credibility information. Two different datasets on health misinformation were used in this study, namely News and Twitter. The effectiveness of the developed method is measured with a Deep learning-based text classifier, Credibility perception method, Hybrid ML approach, Consumer Health Information Credibility Evaluation Technique (CHICET), and Fuzzy Ant Colony Improvement (ACI) system. The developed model attains higher performance in terms of accuracy, True Positive Rate (TPR), and False Positive Rate (FPR) as 0.965, 0.955, and 0.101 with the News dataset. While evaluating the Twitter dataset, testing accuracy, TPR, and FPR measured by the proposed model are 0.961, 0.954, and 0.207, respectively. The results indicate an improvement in terms of credible assessment when compared to existing state-of-the-art models.
telecommunications,engineering, electrical & electronic