REVIEW ON INTRUSION DETECTION SYSTEM BASED METHODS ON KNOWLEDGE DISCOVERY DATA(KDD) DATASET
R Mathiyalagan, Pamela Vinitha Eric · Journal of Critical Reviews · 2020
With the rapid development of network technology, network security has also received more and more attention from researchers of different fields. Early prevention of breach into security systems, Intrusion detection(ID) provides a valuable protection and intrusion detection system provides an added advantage of by decreasing the human resources required for keeping monitoring on the intruders, improving the efficiency of detecting the intrusion, making available information about the security breach that may not have been available without the system, and to help the people learn about these potential risk and also act as an official evidence. Recently, the two major topics that are undergoing a lots of investigations are machine learning and deep learning with more importance given on enhancing the precision of detection classifier. Number of authors and investigators have created intrusion detection system using KDD‟99 datasets. This paper discusses various approaches such as Fuzzy Clustering with Artificial Neural Networks (FC-ANN), K-MEANS algorithm based on Information Entropy (KMIE), Multi-layer Bayesian classifier, Least Square Support Vector Machine (LSSVM), Convolutional Neural Networks (CNNs), Feed Forward Neural Network (FFNN), Recurrent Neural Network (RNN), Deep Convolutional Generative Adversarial Networks (DCGAN) and Improved Relevance Vector Machine (IRVM) developed by different researchers for the detection of intrusion. The gaps pending indicating scope for improvement in these methods are also discussed.