Secured fast prediction of cloud data stream with balanced load factor using Ensemble Tree Classification

Abstract

Cloud infrastructures are used for predicted the data stream with high latency rate on varying load factors with different ensemble models. some of existing stream applications analyze only the temporal relation between data but the Spatio-temporal data information is not processed. For the fast prediction of Spatio-temporal data stream from the cloud infrastructure data distribution, an effective load balancing query processing approach is not spread widely. To achieve load balance statistics on cloud data stream, Ensemble Tree Metric Space Indexing (E-tree MSI) technique is employed and performed with three processes such as scheduling, classification and mapping of cloud data stream for fast effective load balancing. Initially, Fast Predictive Look-ahead Scheduling (FPLS) approach is used to continuously schedule the Spatio-temporal data stream files. The workload of the infrastructure is scheduled in E-tree MSI technique and helps to easily balance the load factor. Secondly, Parallel Ensemble Tree Classification (PETC) in E-tree MSI technique executes the classification operations on cloud data stream. The classification of data stream in Etree MSI technique reduces the overload factor. Finally, bilinear quadrilateral mapping process in E-tree MSI technique linearly predicts the result from cloud data stream storage, with minimal execution time. Experiment is conducted on factors such as linear load balance factor measure, execution time for mapping, and classification accuracy rate.