An Efficient Attack Detection Approach Based on a Proper Feature Selection Strategy

Machine learning techniques are so much used in attack detection strategies for many years. But these methods have issues because they haven’t enough labeled group of data, have much total cost and low levels of accuracy. To get higher accuracy and also get fewer training time, this paper’s goal is using the effective deep learning way based on the feature selection strategies. This method makes the classification of training data using more proper attributes. After training, the model creates good results, which reduces total time for detection and effectively gets better in prediction accuracy. The experimental results show that the proposed method is better than usual machine learning-based intrusion detection methods in terms of simple training, strong adaptability, and accuracy.