Designing a Robust Closed-Loop Intrusion Detection Predictive Model Using Signal Processing Techniques in Cloud Computing Environment
Abstract
As network attacks become more prevalent and complex, it is inevitable to find efficient ways to protect our computing infrastructures. Recently, researchers have begun to harness both machine learning and cloud computing technology to identify threats with reducing the overall computation time of detecting them. The objective of this research is to design an intrusion detection (ID) predictive model to identify abnormal network behaviors (i.e. abnormalities). Advanced signal processing techniques are utilized to design the model. With the model, it would be feasible to protect corporate and government agency's computing infrastructures and data securely. Specifically, this research focuses on 1) extracting significant features that represent the characteristics of abnormal behaviors by applying the signal processing techniques, 2) generating a predictive model to determine and differentiate various attacks (DoS, Probe, and R2L), 3) utilization of a visual analytic tool to identify relationship among the features, and 4) exploring current research trends and directions in network intrusion detection by examining innovative network intrusion detection approaches that utilize both machine learning algorithms and cloud computing technologies. This research is conducted mainly at Bowie State University. The University of the District of Columbia (UDC) joins this project as a sub-awardee.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jul 24, 2017
- Accession Number
- AD1056839
Entities
People
- Dong Hyun Jeong
- S. Choi
- Soo-Yeon Ji