A Survey: An Efficient approach to improve the performance of anomalous detection using adaptive weight butterfly optimization based on traffic prediction

Abstract The most well-known approach to protect a computer network from different malicious activities is to detect the intrusion by utilizing an Intrusion Detection System (IDS). The IDS can be defined as a device or application which plays out the capacity of observing a network. It is necessary to use a systematic and automated IDS creation procedure rather than depending solely on knowledge and experience. This motivates researchers to investigate frameworks for intrusion detection based on data mining. Despite the fact that many algorithms have been developed for intrusion detection, the main challenges highlighted in this study to improve the process of intrusion detection system are controlling the volume of data and removing irrelevant rules. In this research an optimized intuitionistic fuzzy inference system (IFIS) is used to handle the uncertainty in categorizing data packets with unknown traffic patterns. The computer network benchmark dataset Network Security Laboratory Knowledge Discovery in Data Bases (NSL-KDD) is described by the IFIS with the level of membership and non-membership. The difficulty of consistency and ambiguity in evaluating the unknown patterns of irregular packets is solved by using these two degrees to determine the level of reluctance. Conventional Intuitionistic Fuzzy Systems apply the rules produced by the inference engine for categorization without first determining if they have been validated. The classification models accuracy rate is impacted by overfitting as a result. Therefore in the proposed work the main goals of the suggested work are to remove redundant rules, assess rule strength and also to use optimum set of rules. Adaptive weight butterfly optimization (AWBO) is a hybrid task scheduling algorithm which consists of two meta-heuristic algorithms; Monarch Butterfly Optimization (MBO) and Genetic Algorithm (GA). These metaheuristic algorithms generally have applications to unconstrained optimization issues as well as the solution of highly nonlinear systems. In order to solve problems involving global optimization, the Monarch Butterfly optimization Algorithm (MBO) is typically used. Migration and adjustment are the MBO's two primary operators. To tackle with this deficiency MBO is adaptive with genetic operators. A classic metaheuristic algorithm that searches broad areas of candidate space is the genetic algorithm but it needs more function evaluations to get the best possible solution. To propose AWBO which can achieve the optimal solutions for the majority of nonlinear systems as well as unconstrained optimization problems in order to integrate the GA with MBO and solve all of these drawbacks.