Arabian Journal for Science and Engineering (2022) 47:1507–1521
https://doi.org/10.1007/s13369-021-06008-5
RESEARCH ARTICLE-COMPUTER ENGINEERING AND COMPUTER SCIENCE
A Comparative Study on Classifying Human Activities Using Classical
Machine and Deep Learning Methods
Ferhat Bozkurt
1
Received: 10 December 2020 / Accepted: 14 July 2021 / Published online: 26 July 2021
© King Fahd University of Petroleum & Minerals 2021
Abstract
Prediction of human physical activities has become a necessity for some applications that come with the development of
wearable and portable hardware such as smartwatches and smartphones. The task of Human Activity Recognition (HAR)
is to recognize human physical activities, e.g., walking, sitting, and running, using the data collected from sensors, e.g.,
accelerometers and gyroscope. HAR is commonly applied on smart systems, such as smartphones, to serve the understanding
of a user’s behaviors and provide assistance to the user because of the rapid development of ubiquitous computing technology
in recent years. Thus, predicting activities, such as standing, walking, sitting, during the day have become a popular topic in
machine and deep learning. The aim of this study is to predict the user’s activities based on context information gathered by
sensors such as gyroscopes and accelerometers. The conducted classification algorithms extract features from training data
and learn a classification model based on the features to predict activity. In this paper, various classical machine and deep
learning techniques have been studied and compared for human activity recognition. A comparative analysis is performed
between techniques in order to select the classifier with the best recognition performance. Experimental results show that
established Deep Neural Network (DNN) model achieved an accuracy of up to 96.81% and mean absolute error of up to
0.03 on publicly available UCI-HAR dataset. This method has given the best performance between conducted classification
methods in this study to predict human activity.
Keywords Human activity recognition · Smartphone · Sensors · Classification · Machine learning · Deep learning
1 Introduction
Human Activity Recognition (HAR) is a promising technique
which is commonly deployed to identify individual’s activ-
ities using raw sensor data [1]. Recently, HAR is a popular
research topic under the rapid development of ubiquitous
technologies. Classification of human activities has become
one of the essential research topics due to the increasing
dependence of people on technologies that are constantly
developing [2]. In most of the studies, it was aimed to obtain
information about human behaviors with the recognition of
human activities [3]. In activity recognition, data are usually
obtained from outdoor sensor devices or wearable sensors.
Acceleration sensors are preferred more than outdoor sen-
sors with their ease of use, ability to obtain numerical values
B Ferhat Bozkurt
fbozkurt@atauni.edu.tr
1
Department of Computer Engineering, Faculty of
Engineering, Ataturk University, Erzurum, Turkey
in measurement results, affordable costs and use in daily
life [4,5]. Although the data obtained with the outdoor sen-
sors provide information about body movements, they cannot
allow researchers to make detailed examinations. Many suc-
cessful studies have been conducted on the detection and
classification of movements that people perform during the
day by using acceleration sensors [6,7]. After the informa-
tion is collected, activities are recognized by processing them
using machine learning algorithms. Classification of human
activities is used in many areas such as disease diagno-
sis, rehabilitation treatment process, design of smart home
systems, improvement and automation of environmental con-
ditions [8,9].
HAR has been widely applied in many real-world scenar-
ios. It is used as health applications such as the determination
of human activities, calculation of personal daily calories,
analysis of the health status according to the movements of
the person or observing the movements of the elderly peo-
ple for their surveillance. It is also used as human position
tracking and various security applications [10]. The recom-
123