The Role of Machine Learning in Internet-of-Things (IoT) Research: A Review

Abstract

Rapid developments in hardware and software based connected devices with communication technologies introduce Internet-of-things (IoT). In recent year, IoT gained enormous popularity and used vastly in a variety of applications. IoT represents the new forms of communication such as between human and things, and between two things. At present era IoT concepts are highly influence by creating a new dimension in the internet world. Intelligent processing and analysis of big data is the key to developing smart IoT applications. Such applications are logistic, transportation, agriculture, healthcare, and environment. Increasing the use of internet and external factor of environment are caused dynamic change in IoT system. The use of smart environments in the delivery of a pervasive care is the research topic that has witnessed increasing interest in recent years. These smart environments aim to deliver pervasive care through ubiquitous sensing by monitoring the occupant's activities. In order to provide smarter environment, their need to be implement IoT with machine learning. In recent year, machine learning technique have been used widely because of its technologies such that identification, extraction, classification, regression and forecasting. Machine learning exploring historical data from camera and sensors and perform techniques which improve the lifespan of network. In this paper, we build survey on existing research work carried out for various applications of machine learning to IoT. We summarize techniques and tools for IoT and also their benefits and limitations. Using a summarized data we also look for different future challenges. The aim of this paper is to survey different IoT technologies to assist the people to live in a smart environment.

Security aspects in IoT based cloud computing

“…These characteristics include measured service, resource pooling, rapid expansion, network access, and on-demand self-service. Additionally, cloud services are delivered through four deployment models and three service models [7]. The primary objective of cloud computing is to offer various computing services, such as servers, storage, databases, networking, software, analytics, and intelligence, over the internet.…”

Comparative analysis of MapReduce and Apache Tez Performance in Multinode clusters with data compression

“…The term "business intelligence" is employed to describe the analysis of extensive datasets encompassing both social media and business information, which are highly intricate in terms of predicting the dynamic trends in customer demands for various products. Hadoop serves as the ultimate solution for managing big data, incorporating components such as HBase, Hive, R connectors, Mahout, Pig, and Oozie [1][2][3][4][5][6]. These components operate on the Hadoop Distributed File System (HDFS) and the second version of MapReduce, known as "YARN".…”