Circular multi stacked CNN-BILSTM for rumor detection in social networks

Articles in Press

Document Type : Original Article

Authors

1 University of Tabriz

2 Department of Computer Science, University of Tabriz

Abstract

Nowadays social networks have become a widely used platform for communication, entertainment, checking news and events, and more. Although there are many positive aspects such as e-commerce and easy communication through social networks, the existence of rumors is a negative aspect of social networks. With minimal effort, a rumor can quickly spread to many users and lead to unpleasant social phenomena. In previous works, there are some machine learning and deep learning approaches for automatic rumor detection. In this paper, we present a circular multi-stacked CNN-BILSTM model to detect rumors in social networks. From the multi CNN-BILSTMs, a convolutional matrix is created, and then it is made a circular state to the convoluted matrix. This deep learning model only considers the text of social network users and determines whether that text is a rumor or not. We have evaluated the proposed model on the real public PHEME dataset. The proposed model outperforms ten state-of-the-art baseline models in terms of accuracy.

Keywords

Main Subjects

Soft Computing Journal

Articles in Press, Accepted Manuscript
Available Online from 23 July 2024

Files

History

  • Receive Date: 25 July 2023
  • Revise Date: 08 December 2023
  • Accept Date: 19 December 2023

Share

How to cite

Statistics