Certain investigation on adaboost modified classifier for data optimization in internet of things environment

Abstract

The Internet of Things (IoT) network aids in the collection, communication, processing, and application of technology in numerous domains like healthcare, environmental surveillance, transport, manufacturing, and so on. This recent networking paradigm makes use of mobile devices, sensors, actuators, and RFID tags, which are capable of coordinating amongst themselves and utilizing the Internet infrastructure for communication. Wireless Sensor Networks (WSNs) aids in the collection and transmission of data that greatly affect the total performance of the IoT. The large amount of data collected by WSN is either unstructured or semi-structured and is transmitted to IoT for processing. To resolve the storage issues of the huge data generated by IoT, the Hadoop Distributed File System (HDFS) is used to stream the data to user applications as required. It is difficult to accomplish analysis of vast amount of data with existing data processing methods. Thus, in our proposed work data is classified to remove redundant and incorrect data, only relevant data is stored and processed. Various classifiers are used for classifying the data obtained by WSN. In this work, AdaBoost classifier is used for classification of data. To avoid redundant classifiers and also conventional AdaBoost algorithm was built through over consumption of system resources, an ensemble algorithm is proposed in this work. Particle Swarm Optimization (PSO), Genetic Algorithm (GA) and Stochastic Diffusion Search (SDS) with AdaBoost classifier that can reinitialize attributes, thus avoiding reaching local optimum, and optimizing the coefficients of AdaBoost weak classifiers. newline