Comparative analysis of clustering techniques in the internet of things

Abstract

One of the most important topics in the last decade is the Big Data (BD) and how to link it and benefit from its consumption in different fields, included as the introduction in this research analysis of the BD belonging to devices of the Internet of Things. The concept of managing objects and exploring devices is connected to the Internet and sensors deployed in the world, all these devices are pumping a lot of data through the Internet of Things (IoT) into the world. In order to make the right decisions for people and things, BD using data mining techniques and machine language algorithms help make decisions. The Internet of Things that insert large amounts of data need to be studied, analysed and disseminated in order to access valuable, useful and bug-free information for the purpose of making the right decision and avoiding problems. In this paper, two clustering algorithms simple K-means and self-organising map (SOM) in IoT are presented. Next, comparing the clustering models’ output in the IoT data set that improved the SOM is better than K-means, but it is slower in creating the model. Keywords: Internet of things (IoT), big data, machine learning, filtered cluster, K-means, SOM.