A Cloud Based Framework for Identification of IoT Devices at Smart Home Using Supervised Machine Intelligence Model

Abstract

Purpose: Identification of Internet of Thing (IoT) devices in smart home is the most important function for a local server/controller to administer and control the home smoothly. The IoT devices continuously send and receive requests, acknowledgments, packets, etc. for efficient data communication and these communication patterns need to be classified. Design/Methodology/Approach: Therefore, to run the smart home smoothly, in this work a framework using cloud computing is proposed to identify the correct IoT device communicating with the local server based on supervised machine learning. The best supervised machine intelligence model will be installed at the local server to classify the devices on the basis of data communication patterns. Findings/Result: Simulation is performed using Orange 3.26 data analytics tool by considering an IoT devices data communication dataset collected from Kaggle data repository. From the simulation results it is observed that Random Forest (RF) shows better performance than existing supervised machine learning models in terms of classification accuracy (CA) to classify the IoT devices with high accuracy. Originality/Value: A cloud based framework is proposed for a smart home to identify the correct IoT device communicating with the local server based on supervised machine learning. Based on the data communication pattern of the IoT devices, an IoT device is accurately identified. Paper Type: Methodology Paper.