An Efficient and Accurate Detection of Fake News Using Capsule Transient Auto Encoder

Abstract

Fake news is “news reports that are deliberatively and indisputably fake.” News that uses fake information is becoming a threat. It becomes challenging for humans to distinguish between fake and actual news. It has become necessary to detect fake news, which seeks to determine whether a news document can be believed. Detection of fake news faces challenges in accurate classification, making existing detection algorithms ineffective. In these issues, this article uses a novel Adaptive Capsule Transient Auto Encoder (ACTAE) for effectively detecting fake news. ACTAE is a combined approach of a classifier named Capsule Auto Encoder and an algorithm called Adaptive Transient Search Optimization Algorithm. The overall detection process is performed in various stages, including preprocessing, feature withdrawal, feature selection, and classification and optimization of weight parameters of the classifier for better results. The overall process is executed in Python, proving that ACTAE detects fake news with higher accuracy (99%) and lower error rate.