Bio-inspired Hybrid Feature Selection Model for Intrusion Detection

Intrusion detection is a serious and complex problem. Undoubtedly due to large number of attacks around the world, concept intrusion has become very important. This research proposes multilayer bio-inspired feature selection model for using an optimized genetic algorithm. Furthermore, proposed consists two layers (layers 1 2). At layer 1, three algorithms are used selection. The Particle Swarm Optimization (PSO), Grey Wolf (GWO), Firefly Algorithm (FFA). end priority value will be assigned each set. 2 model, Optimized Genetic (GA) select one set based on value. Modifications done standard GA perform optimization fit model. in training phase assign Also, values categorized into categories: high, medium, low. Besides, testing its priority. with high given selected. 2, update may occur selected features calculated F-Measures. can learn modify sets priority, which reflected selecting features. For evaluation purposes, well-known datasets these experiments. first dataset UNSW-NB15, other NSL-KDD. Several criteria used, such as precision, recall, F-Measure. experiments this suggest that powerful promising mechanism system.