Learning Automata with Hyperlink Features for Detecting Venomous Social Trolls on the Social Media Platform

Abstract

Nowadays the widening of illegal activity in the social media, intelligent machinery to detect harmful web pages is required. URL analysis is the best method for detecting phishing and other assaults. Venomous internet robots create fraud posts and start the communication by impersonating a follower or generating several fraud social accounts that are used for venomous purposes. Furthermore, hostile internet robots use shortened harmful URLs in tweets to send queries from online social networking users to venomous servers. As a result, distinguishing harmful internet robots from legitimate users is one of the Twitter network's and instagram's utmost critical responsibilities. To identify harmful internet robots, hyperlink-based data (such as Hyperlink redirect, number of shared hyperlinks, and garbage material in URLs) takes small amount of duration to extract than social chart-based factors (which repeat on the social communication of peoples). A Learning Automata algorithm is used to find the real users of the social media network.