PROJECT TITLE :

Detection of Malicious Social Bots Using Learning Automata With URL Features in Twitter Network

ABSTRACT:

Nefarious social bots create phony tweets and automate their social relationships by impersonating a follower or creating many fake accounts that are used for malicious purposes. Furthermore, hostile social bots use shortened harmful URLs in tweets to send queries from online social NetWorking users to malicious servers. As a result, one of the most significant duties in the Twitter network is identifying malevolent social bots from legal users. Extracting URL-based data (such as URL redirection, frequency of shared URLs, and spam material in URL) takes less time to detect dangerous social bots than social graph-based features (which rely on the social interactions of users). Furthermore, social bots that are harmful are unable to simply alter URL redirection chains. In this paper, a learning automata-based malicious social bot detection (LA-MSBD) algorithm is developed for identifying trustworthy players (users) in the Twitter network by combining a trust computation model with URL-based data. Direct trust and indirect trust are two parameters in the proposed trust computation paradigm. Furthermore, to precisely evaluate the trustworthiness of each participant, the direct trust is derived from Bayes' theorem, while the indirect trust is derived from the Dempster-Shafer theory (DST). Experiments were conducted on two Twitter data sets, and the findings show that the proposed method outperforms existing MSBD algorithms in terms of precision, recall, F-measure, and accuracy.