kENFIS: kNN-based evolving neuro-fuzzy inference system for computer worms detection

Abstract

This paper presents a new kNN-based evolving neuro-fuzzy inference system (kENFIS). The main function of kENFIS is to detect computer worms which possess a constant threat to Internet and have caused a significant damage to business recently. However, kENFIS can be applied to solve complex real-world problems that demand fuzzy rule-based systems able to adapt their parameters and ultimately evolve their rule base. kENFIS partitions the input space into clusters by using a new designed kNN-based evolving fuzzy clustering method (kEFCM) and organizes the rule base using Takagi-Sugeno method. The evolving operation is performed by incremental supervised learning. It integrates the simplicity of k-nearest neighbors (kNN) algorithm with the accuracy of least-square method (LSM) to building up the knowledge-base and learning with a few training examples. The performance of kENFIS has been evaluated and compared with some existing well-known algorithms. Also, its ability to detect worms on-line was tested. The evaluation results demonstrate that kENFIS can be effectively applied in worm detection as well as in other classification problems.