An Efficient Approach for Spammer Detection on Twitter and their Behaviour Analysis

Abstract

Because of social media, internet use, and online marketing, fake news is a serious problem. It seeks to draw in
customers and present false information. Since there is currently no reliable method for differentiating between fake
and legitimate news, deep learning (DL) is a potential academic subject for predicting bogus news. To solve this, an
efficient detection strategy is needed, such as the construction of a Dual-Stage Deep Capsule Auto-Encoder model for
Twitter data. Outperforming existing methods, this work suggests a novel use of deep neural networks for user
geolocation and bogus news detection. Simulation planning assesses the efficacy of proposed methods. A major issue
in public health, politics, the economy, and business is the consumption and dissemination of erroneous or biased
content brought on by the pervasiveness of digital information. Internet users need to deal with this problem. The
method tackles the problem of identifying fake news in contemporary culture by modeling user behavior and
geolocation on social media platforms using psychological elements.