Comparative Study on Fall Detection using Machine Learning Approaches

Conference proceedings article


Authors/Editors


Strategic Research Themes


Publication Details

Author listThienrawit Tongskulroongruang, Paphatchaya Wiphunawat, Wisanu Jutharee, Watchara Kaewmahanin, Teerameth Rassameecharoenchai, Tanagorn Jennawasin and Boonserm Kaewkamnerdpong

Publication year2022

Start page1

End page4

Number of pages4

URLhttps://ieeexplore.ieee.org/document/9795445


View on publisher site


Abstract

Falls are one of the most dangerous problems for the elderly. A reliable fall detection system
can aid in reducing the harmful repercussions of an unintentional fall. The focus of this paper is on a dataset
that includes signals from portable devices. We propose a simple feature selection approach to reduce the number of effective input attributes based on a voting strategy. The dataset is classified into three categories, i.e., pre-fall, fall and post-fall using various machine learning approaches. The computational time is significantly reduced for most of the classification algorithms, and the relative reduction reaches to 66.67% with decision tree algorithm. Classification accuracy can reach as high as 95.17% when using neural network, while the relative reductions on the classification accuracy due to the feature selection do not exceed 1.50%.


Keywords

Fall DetectionFeature SelectionPortable Devices


Last updated on 2023-03-10 at 07:37