Batch Analytics of Data Stream and Outlier Detection in Big Data Using Multiple Classifiers

S.S. Saranya, C. Santhosh and Dr.M. Vijayakumar

The growth of data produced by IoT devices played a vital role. Massive amount of data have been generated by numerous IoT devices in various IoT applications. Whereas Industrial IoT (IIoT) collects enormous data streams continuously from sensors to predict the time to repair or maintain the components of an industry. Several classification techniques provide models of data that predicts the problems or detect the solutions to specific problems by applying the knowledge collected by previous dataset. The highest probability of learning algorithms are used to model the outliers more exactly when input given by classifiers are optimal. Forecasting Future Data Stream (FFDS) algorithm has been proposed to reduce the computing complexity of collected data and predict the future dataset. In batch analytics: collecting, entering and processing a set of data over a time, and then fed into an analytical system. Multiple classification algorithms will enhance the group learning and forecast the time when components of the machine works abnormal.

Volume 12 | 03-Special Issue

Pages: 1496-1500

DOI: 10.5373/JARDCS/V12SP3/20201402