Paper Title
COMPARATIVE STUDY OF SUPPORT VECTOR MACHINE WAR STRATEGY OPTIMIZATION AND CONVOLUTIONAL NEURAL NETWORK SEAGULL OPTIMIZATION WITH FUZZY C-MEANS SPARSE CLUSTERING ALGORITHM AND ADAPTIVE NEURO FUZZY INFERENCE SYSTEM MODELING IN VANET

Abstract
Abstract: The Vehicular Ad hoc Network (VANET) allows for the transmission of secure data to user automobiles. K-means clustering employs the cluster formation paradigm. The cluster head will be chosen using the Tabu Search-based Particle Swarm Optimisation (TS-PSO) technique. Throughput and power efficiency are increased in the proposed technique in an effort to decrease the delay. The adopted blockchain will increase dependability and security. Additionally, an innovative War Strategy Optimisation (WSO) based Support Vector Machine (SVM) model (Optimised SVM) is employed to identify intrusion in the VANET using a trust-based collaborative intrusion detection system. The Vehicular Ad-hoc Network (VANET) ensures the safety of the vehicles during data transmission, albeit occasionally the validity of the data sent can be questioned due to a lack of trust and privacy. The VANET network uses the novel Hybrid Capuchin-based Rat Swarm Optimisation (HCRSO) algorithm for cluster head selection and the implemented Improved K-Harmonics Mean Clustering (IKHMC) model for cluster construction to enable effective throughput, energy use, and delay reduction. Additionally, the VANET model's trust-based collaborative intrusion detection is carried out utilising an optimised (Seagull Optimisation) convolutional neural network (CNN). Blockchain also allows reliability and excellent security. The empirical findings demonstrated that the suggested strategy has higher benefits in terms of malicious node identification, overhead, end-to-end delay, and energy utilisation and is effective to use in cars with limited resources. When there are exactly 100 nodes, CNN-SO has a detection accuracy of 94.3%, whereas SVM-WSO has a detection accuracy of 93.5%. Other methods have less accurate detection. Keywords - Support Vector Machine, Denial of Service, Convolutional Neural Network and VANET etc.