Detecting social media mobile botnets using user activity correlation and artificial immune system

Tweet-Based Bot Detection Using Big Data Analytics

“…A real dataset of benign users and spammers was used, including six newly defined features and two redefined features, and identified three classifiers: decision trees, random forest, and Bayesian networks, for learning. Using Twitter, Al-Dayil et al [77] suggested an identification strategy for social networking related mobile botnets, to identify tweets induced by bots and distinguish those against tweets created by users or by user-approved applications. The proposed approach incorporated the connection between tweeting and user behavior, such as clicking, and an Artificial Immune System (AIS) tracker.…”

Hybrid Machine Learning Techniques Applied in Real Engineering Applications

“…However, we take rich botnet application dataset having C&C ability with state-of-the-art mobile botnet applications sample. Another study by al-detail et al [20] presents Artificial immune system (AIS) with the combination of user activity to detect android botnet from social networking sites as Twitter and discriminate between user-generated tweets or bots caused. Besides, the proposed method worked in 4 steps: (a) capture tweets from twitter (b) relate twitter activity with user's activity (c) create a signature for twitter activity (d) match with existing signatures if it does not match then assumed it legitimate tweet otherwise considers as bots generated tweet.…”