BEST ACCURACY PREDICTION TO NETWORK ATTACKS USING SUPERVISED MACHINE LEARNING ALGORITHM

--To create data for the Intrusion Detection System (IDS), it is necessary to set the real working environment to explore all the possibilities of attacks. The existing system is expensive because of its hardware implementation. Software to detect network intrusion protects a computer network from unauthorized users. The intrusion detector learning task is to build a predictive model (i.e. a classifier) capable of distinguishing between “intrusions” or “attacks”, and “normal” connections. To prevent this problem in network, sectors have to predict whether the connection is attacked or not from KDDCup99 (Knowledge Discovery and Data mining) dataset using machine learning techniques. The aim is to investigate machine learning based techniques for better network connection by finding accuracy prediction results. Machine learning based method is proposed to accurately predict the DOS (Denial of Service) attack, R2L (Remote to User), U2R (User to Root), Probe and overall attacks when compared with the existing supervised classification machine learning algorithms. This accuracy predictor shows that the effectiveness of the proposed machine learning algorithm technique with the existing algorithms are compared for best accuracy by various parameters precision, Recall and F1 Score. KEYWORDS---Dataset, Supervised Machine learning Classification method, Prediction of Accuracy result.