Predict attacker behaviour on IDS with high accuracy using K-nearest neighbor algorithm

The main objective of this work is to predict with high accuracy of the attacker’s behavior pattern by using machine learning methods. The proposed model has improved accuracy, according to the experimental and statical analysis. The study was performed with two machine learning algorithms, K-Nearest Neighbor (KNN) and Decision Tree (DTA). On a dataset of 19, 864 items, the algorithms were implemented, trained, and assessed. The trained and tested dataset has been extracted by performing two iterations on the sample size. Each algorithm has undergone 10 iterations with different test sizes to get different result sets. The G-Power test for machine learning algorithms utilized in this study is roughly 80%. The result sets of the programming experiment have been further analyzed with the statistical tools and observed that the accuracy of the KNN is 99.76, while the DTA is 98.84, according to the testing data. Utilizing independent samples t-tests, the statistical difference is p