ABSTRACT
In traditional attack detection, the popular method is to establish the normal behavior pattern of the network data stream. This chapter utilizes Recurrent Neural Networks (RNN) to analyze and represent the original attack features and depth features and Support Vector Machine to detect an attack with the depth features. Network space is open and makes life convenient. However, from many types of network security event reports, security vulnerabilities and risks are emerging in an endless stream. There are many types of attack detection methods, such as feature matching, expert system, state transition, fuzzy reasoning, data statistics, pattern recognition, machine learning, and deep learning. Compared with the traditional attack detection methods, our method can take full advantages of feature information with RNN, which is an artificial intelligence method. Attack detection has a high practical significance, and depth feature representation is a hot research in feature processing.
