Neural Network Training by Selected Fish Schooling Genetic Algorithm Feature for Intrusion Detection
Nomaan Jaweed Mohammed · International Journal of Computer Applications · 2020
In an ever-growing world of internet users, network security has become an important aspect of today's digital age.Due to a multitude of users accessing the internet for a plethora of reasons, it has become imperative to identify an appropriate and safe network for which, an Intrusion Detection System (IDS) solution has been proposed.The proposed IDS solution utilizes Fish Schooling Genetic Algorithm and an error backpropagation neural network.The genetic algorithm has been used for detecting the good feature set from the training dataset and the selected good features train the neural network.This combination of genetic algorithm and Neural network increases the detection accuracy of intrusion with a lesser number of training features, and the reduction of the feature set increases the learning accuracy of neural networks for intrusion detection.This experiment was done on a real dataset and the obtained results are better than the previous works done on different parameters that are highlighted in Section II below.