Reducing Unwanted Service Utilization Based On Invariant Slide Window Feature Selection Using Feed Forward Nueral Network in Cloud Server Utilization

Abstract

Cloud load balancing is distinct as a process of dividing loads and computing possessions in cloud computing. Businesses can manage assignment strains or request stresses by distributing resources across a large number of computers, networks or servers, mainly the traffic in cloud is a big problem approach in service utilization. Carrying the time factor is nondeterminant to improve the quality of the service. Cloud load balancing involves keeping workload traffic and requests distributed across the Internet. This reduces the non-related variance features by using Invariant Slide Window Feature Selection (IVSWFS). The selected features difference be trained with spider optimization using feed forward modified Support vector machine (FFSVM) algorithm. This predicts the traffic flow by class by variance level as well classification in precision, recall rate compared to other system.