Intrusion Detection System Using K-Means Based on Cuckoo Search Optimization

Abstract

Recently, data protection is very important with the technological and digital revolution, as a vast amount of data is generated from different networks. It was found that the Intrusion Detection System (IDS) is probably the best option because of its ability to differentiate between threats that occur inside or outside a public internet. Cluster analysis is a common method of data mining and is characterized as the grouping of similar data. One of the clustering algorithms for clustering numerical data is K-Means. The K-Means Algorithm features are simple to implement and large amounts of data can be handled efficiently. Natural optimization algorithms have recently been combined with clustering algorithms in order to reach the best global solution. Algorithm for optimization search in Cuckoo is a recent meta algorithm for heuristic optimization. The intelligent behavior of the cuckoo is based on this algorithm. Cuckoo Search Optimization (CSO) and the K-Means clustering algorithm are combined in this paper to achieve the optimal solution globally. Different data sets are evaluated and results are compared with those of the clustering algorithms based on optimization.