kumoh national institute of technology
Networked Systems Lab.

Fabliha Bushra Islam, Cosmas Ifeanyi Nwakanma, Mareska Pratiwi Maharani, Jae Min Lee and Dong-Seong Kim, "Activity Pattern Recognition using Machine Learning Classifier for IoT-enabled Smart Home", 2020 Fall Conference on Korea Institute Of Communication Sciences(KICS), Nov. 13, 2020, Online, South Korea(N8&N12)
By : Bushra
Date : 2020-10-20
Views : 67

Abstract—This paper presents a classification analysis based on Machine Learning (ML) algorithms for recognizing jumping activity in smart home scenario. These activities are captured by vibration sensor G Link 200 in real time. After collecting data set, a sufficient number of ML classification techniques are applied to experiment the accuracy based on activity pattern recognition. Two classifiers Weighted K-Nearest Neighbour(KNN) and Kernel Naive Bayes(KNB) algorithms outperformed other classifications with the highest accuracy of 99.2% and 98.7% respectively. This solution can be efficient for observing anomalous behaviours for Industrial IoT networks, and smart factories, smart home management to restrain future inconveniences.

Index Terms—Machine Learning (ML), G Link 200, Smart Home, Internet of Things

(Total:277 articles / page:1/28 )