Adaptive Salp swarm optimization algorithms with inertia weights for novel fake news detection model in online social media

Abstract

Recently, social media are the most popular way of consuming news for people due to their fast, low cost, and easy accessibility. Unfortunately, in order to provide financial, political, or personal interests on social media, a large amount of fake news is intentionally produced that contains false information. Although fake news detection is a very important problem to avoid negative effects, efficient studies on this issue are limited. More efficient models are required in order to obtain better solutions with respect to different metrics for fake news detection. In this paper, a novel model was proposed that uses optimization methods for fake news detection. In addition, an improved Salp Swarm Optimization (SSO) based on a nonlinear decreasing coefficient and oscillating inertia weight was proposed to find the best optimum solution for fake news detection for the first time. The standard SSO, Grey Wolf Optimization (GWO) which is one of the most recent swarm intelligence algorithms, and two new adaptive SSO methods were modeled to detect fake news for the first time in this study. These methods were tested over four different real-world fake news data sets to verify the performance of the algorithms proposed in this paper. Furthermore, Friedman test was conducted to distinguish the differences among these methods. The obtained results prove that the proposed new model is significantly superior to standard SSA and GWO on the real-world fake news data sets.