A prediction system of Sybil attack in social network using deep-regression model

Sybil attacks have grown prevalent in Twitter and other social networks owing to the increase in the number of users now found on these very popular platforms. Sybil accounts are thus escalating in number with many of the operators of these accounts always adapting their techniques to evade detection. Such is the complexity of many Sybil profiles that most Sybil detection techniques are no longer very effective in preventing and controlling their activities. For this reason, it is vital that the detection techniques are optimized with fresh data with an aim of improving the strategies against the ever evolving Sybil operators. In this paper we introduce a prediction system that can be leveraged in the manipulation of deep-learning solution model hence solving the problem of Sybil attacks on Twitter. Our proposed system includes three integrated modules, namely, a data harvesting module, a feature extracting mechanism, and a deep-regression model. All of these modules function in a systematic form to analyze and evaluate user's profiles on Twitter. The proposed model looks to deliver this kind of optimization and it has proved to deliver an accuracy of up to 86% when fed with unclean and noisy data.