Application of Deep Learning for Database Intrusion Detection

In this paper, we have suggested a deep learning model aimed at effective detection of malicious transactions in a database system. This method focuses on exploiting the user normal behavior, data dependencies, and data sensitivity of a transaction to predict intrusions. Currently, we have used different kinds of neural networks according to their strengths of predicting the intrusion according to the type of data such as sequential or featured data. For experimental evaluation, we have used a recurrent neural network for sequence data and feed-forward with back propagation for other attributes, together creating a hybrid deep learning model which works effectively to predict the database intrusions.

[1]  Yi Hu,et al.  A data mining approach for database intrusion detection , 2004, SAC '04.

[2]  Shamik Sural,et al.  Two-stage database intrusion detection by combining multiple evidence and belief update , 2013, Inf. Syst. Frontiers.

[3]  Li Deng,et al.  Sequence classification using the high-level features extracted from deep neural networks , 2014, 2014 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[4]  Ajith Abraham,et al.  Modeling intrusion detection system using hybrid intelligent systems , 2007, J. Netw. Comput. Appl..

[5]  Yuefei Zhu,et al.  A Deep Learning Approach for Intrusion Detection Using Recurrent Neural Networks , 2017, IEEE Access.

[6]  Lilly Suriani Affendey,et al.  Intrusion detection using data mining techniques , 2010, 2010 International Conference on Information Retrieval & Knowledge Management (CAMP).

[7]  A. John,et al.  Survey on data mining techniques to enhance intrusion detection , 2012, 2012 International Conference on Computer Communication and Informatics.