Summary With expanding nature of Internet of Things (IoT) based solutions in various application areas, the threat of cyber attacks is looming large. Often for resource constrained IoT nodes, implementation of machine learning based intrusion detection system is not practical, as the feature set to be processed is large in size. In this article, random forest based intelligently initialized hybrid binary particle swarm optimizer (PSO)‐gray wolf optimizer (GWO) is proposed to reduce the dimensions of dataset. The proposed algorithm exploits the concept of relative weights of the leader wolves of GWO to update the positions of particles in PSO. A novel fitness function is also introduced, that includes the key performance metrics for measuring the classification efficiency. Also, a new performance metric is proposed here for commensurate comparison with other related works, which renders the task of comparison effective. For experimental evaluation, the historical dataset NSL‐KDD and the more contemporary DS2OS dataset are both taken into consideration. The proposed work attains accuracy up to 99.61% for NSL‐KDD dataset and 99.79% for DS2OS dataset. The outcomes are better than majority of algorithms and recent related works. Most notably, for increasing the accuracy, the number of features are not compromised and are even reduced to 8 and 5 features, respectively, for NSL‐KDD and DS2OS.

Summary The current researches are primarily focused on optimizing the available classification methods in support vector machines (SVMs). The basic idea of this article to highlight the fundamentals of the proposed SVM variants and their optimization techniques. In this article, we have collectively identified the major issues in different variations of SVMs. This article has discussed a detailed explanation of the optimization and advancement of SVM and its kernel variants. The study elaborates on certain SVM optimization issues like model selection, novelty detection and so forth. In contrast, it discussed the identification of problems raises after every optimization and advancement of SVM. This is the first novel attempt to explain the optimization raises for SVMs. This article is about the majority of the necessary advancements in the optimization and cost‐effectiveness of computational algorithms of SVM with their detailed computational analysis. The approaches of such study helps to understand the optimization challenges in SVM.

Abstract The Internet of Things (IoT) is the domain of interest for researchers with exponential growth in technology. This article develops an authentication protocol based on the Chebyshev polynomial, hashing function, session password, and encryption. The proposed authentication protocol is named as proposed Elliptic, Chebyshev, Session password, and Hash function (ECSH)‐based multilevel authentication. For authenticating the incoming user, there are two phases, registration and authentication. In the registration phase, the user is registered with the server and authentication center (AC), and the authentication follows, which is an eight‐step criterion. The authentication is duly based on the scale factor of the user and server, session password, and verification messages. The authentication at the eight levels assures the security against various types of attacks and renders secure communication in IoT with minimal communication overhead and packet loss. The performance is analyzed using black‐hole, and denial‐of‐service (DOS) attacks with 50 and 100 nodes in the simulation environment. The proposed ECSH‐based multilevel authentication acquired the maximal detection rate, PDR, and QoS of 15.2%, 35.7895%, and 26.4623%, respectively, for 50 nodes with DOS attacks. By contrast, the minimal delay of 135.922 ms is acquired in the presence of 100 nodes and DOS attacks.

Abstract The increased usage of online social networks in recent days attracts the attention of researchers. That makes analysis of social networks a significant concern. Link prediction is one of the problem of social networks, where links are predicted between users by analyzing the relationships between the nodes. As the size and usability of the social network grows, the accurate prediction of connections with limited public information is a challenge. This article introduces a novel approach for predicting missing links using a nature‐inspired method, the gravitational search algorithm. We selected seven real‐world networks and five widely used algorithms and also considered the sparsity of networks for the experimental evaluation of the proposed approach. The proposed GSA‐driven algorithm has shown a better or comparable AUC result than the other standard algorithms on these networks.