Support Vector Machine (SVM) Based on Wavelet Transform (WT) for Intrusion Detection System (IDS)

In this paper, SVM based DWT used to detect the IDS problem. SVM have many applications in different real live problems. The proposed method consists from two-part DWT and SVM. The DWT used to extract best and sensitive features from training data. The extracted features from DWT become input to the SVM and the SOFTMAX trained to classify the input data into two labels there is attack or not. The experiment was implemented using MATLAB2016 on a dataset consist from 175,341 instance, each of these instance consist from 42 features and validated using 82,332 instance. The proposed method is first time used to detect IDS problem and Produces 95.92% accuracy when validated by using UNSW-NB15 dataset. The experimental results show the proposed method presented satisfactory results when compared with best results obtained in this field.