Social Engineering Detection Using Neural Networks

Social Engineering (SE) is considered to be one of the most common problems facing information security today. Detecting social engineering is important because it attempts to secure organisations, consumers and systems from attempts to gain unauthorized access or to reveal some secrets by manipulating employees. The aim of this work is to introduce a new technique for detecting social engineering using neural networks. In this work we have used benchmark data and developed a new technique to extract features that can be used for neural network testing and training. Initial results are encouraging and indicate that machine learning can add an extra layer of security to protect individuals and organisations from social engineering attacks. Future work includes expanding the data set to include additional attack scenarios and benchmark data.