Paper Info
Title
Exploiting Behavioral Differences to Detect Fake Ne
Abstract
Online social platforms have become the most influential media and their impact will be far greater in the highly connected and super-efficient smart cities. The speed, reach and sheer volume of digital media pose a global challenge to combat fake news. There is an urgent need to build resilience in a post-truth era. Misinformation gnaws social cohesion and erodes the trust of the citizens. This study seeks to identify the key differences in the traits between fake news and normal information in tweets and present two case studies to showcase two such features, namely, user sentiments and spread pattern. We then we propose an AI-based system using Autoencoder and Recurrent Neural Networks to detect fake news in Sina Weibo. This Proof-of-Concept (POC) can achieve a reasonable accuracy and F1 score and also proves its applicability to other online social platforms. The proposed POC is especially useful for governments, companies and other organizations to identify such misinformation as early as possible so that immediate actions can be taken to minimize the potential negative effect. It can also be deployed for use by social media platform users.
Year
DOI
Venue
2018
10.1109/UEMCON.2018.8796519
2018 9th IEEE Annual Ubiquitous Computing, Electronics & Mobile Communication Conference (UEMCON)
Keywords
Field
DocType
Fake news detection,Feature selection,Autoencoder,Recurrent neural network
Psychological resilience,Data science,F1 score,Social media,Autoencoder,Feature selection,Computer science,Misinformation,Recurrent neural network,Human–computer interaction,Digital media
Conference
ISBN
Citations 
PageRank 
978-1-5386-7694-3
0
0.34
References 
Authors
0
7
Name
Order
Citations
PageRank
Weiling Chen100.34
Chenyan Yang200.34
Gibson Cheng300.34
Yan Zhang4423.72
Chai Kiat Yeo562472.12
Chiew Tong Lau640635.82
Bu Sung Lee745235.22