Abstract: Social media are websites that provide a network of people channels to make connections. An example of the media is YouTube that connects people through video sharing. Unfortunately, due to the explosive number of users and various content sharing there exist malicious users who aim to self-promote their videos or broadcast viruses and malwares. Even though the detection of malicious users is based on various features such as content details, social activity, social network analysing or a hybrid of features, the detection rate is still considered low (i.e., 46%). This study proposes a new set of features which is based on edge rank concept that focuses on affinity, weight and decay. The research is realized by analysing a set of YouTube users and their shared video prior to classifying the users using seven classifiers. Evaluation is performed by comparing the classification results of the proposed features against the existing feature set. Experiments showed that 86% of the classifiers obtained better results when using the proposed feature set as compared to using the existing features. The average classification accuracy is at 95.6%. Such a result indicates that the proposed work would benefit YouTube users in obtaining the required multimedia content and creating trust among users. In addition, system resources can be optimized as malicious accounts do not exist.
Omar Hadeb Sadoon and Yuhanis Yusof, 2017. Detecting Malicious User in YouTube Using Edge Rank Based Feature Set. International Journal of Soft Computing, 12: 7-12.