DEVELOPMENT OF MULTI LEVEL INTRUSION DETECTION SYSTEM FOR CLOUD BASED LOG MANAGEMENT USING MACHINE LEARNING TECHNIQUE

Subject Area: COMPUTER SCIENCE

This paper presents development of a Multi Level Intrusion Detection System (MLIDS) for cloud-based log management using machine learning technique. The research methods are data collection, data processing, multi layered neural network, threat detection algorithm, threat control algorithm and multi level intrusion detection system developed with the threat algorithms. The algorithms were modeled using structural methods which engaged Universal Modeling Language (UML) diagram for the system design. Pseudocodes were also used to present the generated algorithmic outputs which were used to model the MLIDS. The system was implemented with Simulink, tested and validated with Mean Square Error (MSE) and Regression (R). The result of MSE is 2.47e-05Mu and R= 0.99445. The implication of the results showed that the new algorithm developed was able to correctly monitor, detect and prevent threat penetration to cloud-based server. A comparative analysis was also conducted with other threat detection algorithms and from the result, it was observed that the performance of the new system was better due to its multi layered configuration of neurons to enhance data processing and computation in the hidden layers, then right choice of activation used, training algorithm adopted and also the quality of data used to train the neural network and achieve the threat detection algorithm. These features of the new algorithm make it to standout from the others with better performance.