Pattern Recognition of Human Activity Based on
Smartphone Data Sensors Using SVM Multiclass
Alman
1
, Armin Lawi
2
, Zulkifli Tahir
3
{alman16p@student.unhas.ac.id
1
, armin@unhas.ac.id
2
, zulkifli@unhas.ac.id
3
}
1
Department of Electrical Engineering, Universitas Hasanuddin, Indonesia. 92119
2
Departement of Computer Science, Universitas Hasanuddin, Indonesia. 92119
3
Department of Informatics Engineering, Universitas Hasanuddin, Indonesia. 92119
Abstract. Mobile devices are increasingly sophisticated while smartphones continue to
make the latest generation that immerses the supporting tools needed in everyday life such
as cameras, GPS, Microphones, and various sensors such as light sensors, a direction
sensor, acceleration sensor (i.e., accelerometer) and the gyroscope sensor. This study aims
to classify human activities from the accelerometer and gyroscope sensors on a Sony z3+
smartphone. To implement our system, we collect labeled accelerometer and gyroscope
data from eight users when they carry out daily activity. Every activity was recorded for
22 seconds, total data that we use every activity is 2000 data with the total amount of data
is 16000 data. This data we classify using the Multiclass Support Vector Machine (SVM)
method reaches 97.40% accuracy using a 70% ratio as training data and 30% as test data,
the classification process takes 5 seconds to classify the data.
Keywords: Activity, Recognition, Accelerometer, Sensor, Support Vector Machine.
1 Introduction
Mobile phones only can be used for call and sending a message. But as time goes by a mobile
phone can be used to perform various tasks, for example for typing, internet, chatting, editing
photos, sending emails, playing games, navigators, compasses, and we even use smartphones to
monitor our health through recording activities, finding our location on maps, watching videos,
and buying products on the Internet[1]. People-centric sensing says[2] using a smartphone as a
sensing. Smartphones are now popular and its a need for people in recent years[3]. At first, the
mobile phone components were very simple, but now a mobile phone has various sensors
embedded in it.
With this sophistication, today's mobile phones are called smartphones, the smartphone can
be called because it has a variety of sensors that make a cell phone has many features. Embedded
sensors on Smartphones can be utilized now. One of the advantages of introducing this activity
is able to predict the potential for a user's fall[4] and it immediately sends notifications to the
right people (i.e., family or medical staff)[5] and also it can provide information even though
the user is sitting, walking, running, standing, sleeping, getting up, going upstairs and going
downstairs. From the results of reading the sensor information can be used in this study to design
the system in classifying the results of sensor data
This sensor output is in the form of data where the data generated has hundreds or even
thousands of records. Of course, the data cannot be read or understood only by looking at it or
reading it directly. To find information that is not known before, potential[6] knowledge that
ICOST 2019, May 02-03, Makassar, Indonesia
Copyright © 2019 EAI
DOI 10.4108/eai.2-5-2019.2284606