BIFDeM: A BIO-INSPIRED NOVEL MECHANISM FOR FAKE NEWS DETECTION

ABSTRACT

The existence of the internet and its ease of accessibility has made information sharing quite easier and faster, but these advantages come with a corresponding consequence as it aids in the spread of fake and non-factual news. When a news lacks integrity or its misleading, it is known as fake news. The spread in fake news therefore calls for an accurate model for detecting whether a news is fake or not. Till date, the field of fake news has seen significant research in the development of mechanisms and models for abating the spread of fake news mostly applying the traditional machine learning algorithms.

However, bio-inspired algorithms which mimics the behaviour of the biological immune system (a network of biological processes which protects its host from foreign pathogens), have shown excellent performance in spam detection and as such can be applied to fake news detection since they are both classification problems. The Negative Selection Algorithm (NSA) is one of such bio-inspired algorithms that is based on the principle of self and non-self and has proven to be well suitable for classification problems. Thus, our proposed model, BIFDeM (Bio-Inspired Fake news Detection Model) aimed at distinguishing accurately fake news from real news using linguistic features.

Keyword: Fake news; Negative selection; Computer security; Algorithms, Detection Model.