Document Type : Research Article

Author

• L K Suresh Kumar

Abstract

Currently, problems such as a slow processing speed, a low detection rate, a high false-positive rate, and a big feature dimension are all part of intrusion detection. To handle these issues, Iterative Dichotomiser 3 (ID3), Principal Component Analysis (PCA), and Back Propagation (BP) algorithms are available. Through a lower false-positive rate, and a higher detection rate, the research-based intrusion detection model ID3-PCA-BP increases the processing speed of intrusion detection systems. To decrease the overall data volume and accelerate processing, ID3 is used to initially differentiate the data. Differentiate ID3s and keep the temporary training sample set for intrusion data in order to retrain and optimize the ID3 and BP, treat the ID3 judges as standard data, and delete the added intermediate data. Then, we should reduce the dimension of the data using PCA and then introduce the data to BP for secondary discrimination. However, the ID3 algorithm operates a shallow structure in order to prevent an extreme quantity of intermediate numbers from being analyzed as intrusion data. As a result, further BP processing cannot effectively raise the accuracy. BP accelerates data processing by employing the ReLU activation function from the simplified neural network calculation approach and the faster convergence ADAM optimization algorithm.

Keywords

• Intrusion Detection
• Machine Learning
• Iterative Dichotomiser 3
• Principal Component Analysis
• Back Propagation