10 III March 2022

https://doi.org/10.22214/ijraset.2022.40905
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue III Mar 2022- Available at www.ijraset.com

Smartphone User Behaviour Predication Using AI

Mayur Ramesh Chavan1, Akshay Avadhut Kulkarni2, Mohammad Kaif Shakil Mulla3, Ashish Hanmant Mahadik4,
Omkar Jagnnath Waghmare5, Dr. S V Balshetwar6
1, 2
Dr. Daulatrao Aher College of Engineering Karad

Abstract: There are lots of Smartphone users data floating around which is not efficiently used for improving user experience
and which can be used to provide a better user experience or enhance the current one. Enhancing user experience is a challenge
as there are multiple competitors. A persons time spend on their Smartphone is an vast amount of time and everyone is trying to
reduce this time by implementing efficient ways of android user experience. Starting with our work, which will focus on this user
floating data? There are important segments of this user data one of them is facial data because humans are used to taking in
non verbal cues from facial emotions. And also we are focusing on factors like user surrounding sound, user’s locations and
patterns of application used history.
I. INTRODUCTION
As in the 21st century Electronics devices are conquering the whole world and use of Smartphone is increasing day by day in which
most of the time spent on smart phone is less productive, whereas it leads to wastage of time. So while studying an average
Smartphone user we found that the Smartphone can cause two different problem, where first one shows a negative and second one
shows positive impact of Smartphone . Coming to the first problem shows us the deficiency.
So now we will discuss the first problem and further continue to explain second problem. There are lots of Smartphone user’s data
floating around which is not efficiently use for improving user experience and which can be used to provide a better user experience
or enhance the current one. Enhancing user experience is a challenge as there are multiple competitors. A persons time spend on
their Smartphone is an vast amount of time and everyone is trying to reduce this time by implementing efficient ways of android
user experience. Overusing pattern of Smartphone involves a tendency to check notifications all the time. Such behavior pattern can
induce “reassurance seeking” pathway which broadly includes symptoms such as loneliness, low self-esteem, depression, and
anxiety. Excessive use of Smartphone may also affect sleep patterns by reducing rapid eye movement sleep, slow-wave sleep and
consequently causing sleep.

II. LITERATURE REVIEW

Coming to the first problem discussed above that is the overusing pattern of Smartphone involves a tendency to check notifications
all the time. Such behaviour pattern can induce many health problems. After studying these problems we found that there can be a
single solution, which will be focusing on prediction of an average Smartphone user behaviour. As this can be a challenging task
because there are varieties of peoples from different age group, different background, etc.
So we decided to focus on a specific group of people which are “Students”. As student are the most suffering from Smartphone
addiction as they are in there early years of education life. Student may feel overwhelmed, curious, lonely, bored, stressed,
depressed or even anxious about something. Once they enter in virtual world of Smartphone applications, they might feel to forget
worries and feel better. But this can further lead to addiction, as student can make Smartphone as comfort zone and spend more and
more time on Smartphone. Smartphone is a device that brings the world closer to us but excessive use of it can disturb real world
connections.
As we have studied problems related to Smartphone user behaviour found some solutions and these solutions are highlighted are in
our research. In [1] Ankita Kanhangadand and et.al.in there paper had taken the data from the sensor like orientation sensor,
accelerometer sensor and gyroscope sensor which are built-in sensors. And from these dataset taken from the various sensors they
are going to only predict the human behaviour. The analysis of human behaviour has been proven to be effective in various
applications including biometricbased user authentication, smart spaces, human-machine interactions, physical activity recognition
and surveillance. The advantage of this thing is human behaviour is captured unobtrusively without requiring a conscious effort on
the part of user. This study will help us to explore techniques that can be used to predict human behaviour.
As per [2] Subrata Tikadar and Samit Bhattacharya are it is very important to know user behaviour to design and built effective
interaction system, tools or application.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 1
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue III Mar 2022- Available at www.ijraset.com

The behavioural study not only helps to assure the success of any design or product but also helps other researchers from various
related areas. They have systematically collected and analyzed the behavioural data for Smartphone usage by 1711 students of 188
academic institutions throughout India.
They have also observed students behaviour on Smartphone usages both inside and outside the classroom. They have also found
dissimilarities and are also expecting that the findings of the study will help many researchers from various fields including HCI,
Mobile HCI, Behavioural Science, Psychology, and Education. This study will help us to know importance of prediction of human
behaviour and explore its depth knowledge.
As per [3] authors have here presented the article, which present the frameworks for mobile audience measurements, for collecting
data at the point of convergence - devices.
This paper compares the presented framework to alternative methods of mobile user research, and identifies the unique advantages
of on-device measurements along with the key weaknesses. This study shows us that there are various techniques of mobile
audience measurement from various aspects through a Smartphone.
In[4] Ruxia hong has focused on the research of mobile user behaviour based on big data and it has become one of the hotspot in the
field of internet, Every internet user leave their footprints.
It is based on the analysis of the characteristics of mobile user's big data and hadoop system, an analysis model of mobile user's
behaviour pattern based on big data is constructed, it includes data acquisition module, data pre-processing module, user behaviour
analysis module, application of mobile user behaviour model and data visualization module, and the function of each module is
explained in detail. This study shows us that big data can play a measure role to analyse data sets and also that internet can be a vast
plays to explore Smartphone user data.
As per [5] author has given the model of machine learning for activity recognition and authentication of Smartphone user.
As per Authors technological advancements have made Smartphone’s to provide wide range of applications that enable users to
perform many of their tasks easily and conveniently, anytime and anywhere. As per their study many users are tend to store their
private data in their smart phones.
Since conventional methods for security of Smartphone’s, such as passwords, personal identification numbers, and pattern locks are
prone to many attacks, this research paper proposes a novel method for authenticating Smartphone users based on performing seven
different daily physical activity as behavioural biometrics, using Smartphone embedded sensor data. This authentication scheme
builds a machine learning model which recognizes users by performing those daily activities. This study shows us that Smartphone
sensors can be combined and used for user identification to increase security measures.
As per [6] Balaji Balasubramanian and Pranshu Diwan et. al. signify in their paper that human beings relay a lot on non-verbal
communication and facial emotion in large. In this Paper they cover the dataset and algorithm that are used for Facial Emotion
Recognation(FER). And the algorithms range from simple Support Vector Machines (SVM) to complex Convolutional Neural
Network (CNN). This paper shows us importance of facial emotion and how it can play a measure role in our work since they have
also represented various techniques for recognising it.
As per [7] Richard Han, Mahnaz Roshanaei and Shivakant Mishra all authors present in this paper have formed a group of 20
peoples for a research that put in the picture how every individual from the group reacts in different scenario that relates with respect
to their Smartphone. They experimented most of the possible scenarios and prepared data sets to be tested with machine learning
algorithms for the better accuracy to knowing their behaviour.
As per [8] Natasha Jaques, Sara Taylor, Asaph Azaria and et.al. focused in this paper that an average teenagers happiness depends
on some of their close things that also includes there Smartphone’s and other things like sleep. These things can affect measure
things in their life and can also result in their personal losses and they also found 70% accurate module that shows this behaviour.
This study can help us with our research to know about this factors and there consequences.
As per [9] mobile phones are equipped multiple sensors from which there are new opportunities to analysis user’s daily behaviours
and also how truly intelligent personal devices are.
They have proposed a MAST (movement, action & Situation over Time) model to explore along this direction and identify key
technology required. And also they have found an idea of reducing power consumption for mobile phones with the help of phone-
cloud collaboration model. This study shows us a new model known as MAST and it can be very valuable in our survey.
As per [10] Smartphone have lots of resources that can be utilized to enhance user behaviour analysis. Study shows that how
Psychological Science can be used to study user behaviour, it gives an idea about various opportunities for analysing Smartphone
and it can be helpful in our work.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 2
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue III Mar 2022- Available at www.ijraset.com

Below Table1.1 shows comparison of different techniques used to analyse Smartphone user behaviour

Table 1. Comparison Table

Sr. Paper Name Publisher Techniques Merits Demerits

No
.

1 Smartphone usage Qian He, Here 1.With the help of The paper is only
contexts and Emmanuel O. logistic logistic Regression they limited to students
sensible Agu Regression are able to classify user and student age
patterns as technique is context variable such as group only.
predictors used location, time and app
of future usage.
sedentary 2.The paper also shows
behaviours that users are very
sedentary or not, means
in other words how much
user is spending time as
seated.

2 Smartphone app Raihana Here predicting 1. By This paper only

usage as a Ferdous, Venet a stress level of understanding the show smartphone
predictor osmani, Oscar user patterns of app usage user stress levels
of Mayora based on and and it is not enough
perceived stress smartphone investigating relationship to predict it next
levels as app usage of these pattern the move.
workspace. technique is perceived stress level
used within the workspace
context.
2. The result they
have achieved is average
of accuracy of 75% and
precision of 85.7% can
be used as an indicator of
over all stress level in
work environment.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 3
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue III Mar 2022- Available at www.ijraset.com

3 Machine Learning S. Sareh Here normal 1. This paper tend to The proposed way
models for Ahmadi, Sherif technique for provide a better way to of security requires
activity Rashad, Heba security of increase security of users 7 daily activates
recognition and Elgazzer smartphone data. and it leads to more
authentication of and 2.With the help of seven time consumption.
smartphone users. conventiona different physical
l method for activities like behaviour
same is biometrics, using
used. smartphone embedded
sensor data security is
provided in the sense of
passwords,
personal identification
numbers, and pattern
locks etc.
4 Artificial Chia-Liang Here four 1. With the help of This paper only
Intelligence Chen, Fu-Ming supervised logistics regression, and show the best
and Mobile Huang, Yu- machine support vector machine algorithm for User
Phone Sensing Hsin Liu, DaiEn learning automatic activity Activity
based User Wu technique is classification model is Recognition but not
Activity used and created. tend to give an
Recognition various 2. They have patter or prediction.
classificatio evaluated the
n models are prediction performance
made and the results of these
experiments shows that
under specific
acceptance of accuracy
and minimum model
training time, the
decision tree algorithm
creates the best model.
5 Identifying Mohammed A. Here also 1. The maturity in sensor This paper tends to
smartphone users Alqarni, Sajjad they had chips and machine use more
based Hussain used the learning algorithms physical activities
on how they Chauhdary, technique of provides a better solution than use of
interact with their Maryam machine for authentication smartphone itself.
phones Naseer Malik learning problems based on
various behavioural biometrics,
algorithms, which aims to identify
Gesture the behavioural traits that
recognition and a user possesses,
behavioural

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 4
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue III Mar 2022- Available at www.ijraset.com

biometric is such as hand movements

also used. and waving patterns etc.
2. Their research study
aims to provide a solution
for passive and continuous
authentication of
smartphone users by
analysing their activity
patterns when interacting
with their phones.
6 Is Smartphone Liette Here they 1. As per the investigation The paper show the
usage is truly Lapointe, have used a i.e. the 11 depth study that show if a
smart? A Camille grounded interviews and answers to person is addicted
Qualitative Boudreau- theory to 183 exploratory questions to smartphone or
investigation of IT Pinnsonneault, report the they have revealed out of not but it does not
addictive issac vaghefi result of four smartphone user specifies it types or
behaviour addictive profiles two of these are any further details.
smartphone exhibiting addictive
usage. behaviour.

7 A Smartphone user Peng Dai, Shen Here they 1. With the help of The study
activity Shying ho have made Activity Prediction conducted through
prediction their own Framework they are giving this paper is
framework technique/al prediction of next day based on old datasets
utilizing partial gorithm to find behaviour of same user that will not be
repetitive and a based on weighted sum of useful that much.
landmark repetitive most similar behaviour
behaviours behaviour of vectors related to landmark
smartphone behavour of next day
user. behaviour.
2. With the help of
arbitrary call activity,
voice call activity, short
message activity, media
consumption and app
usage datatypes, extensive
experiment are carried out
using nokia mobile data
challenge (MDC) dataset
to demonstrate
the
feasibility of their
proposed approach.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 5
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue III Mar 2022- Available at www.ijraset.com

8 Usage Katsuki This is the 1. They are The paper show

Prediction and Yasudomi, study paper focusing on app restriction study that only
Effectiveness Toshitaka which function, which is one of focuses on one
Verification of Hamamura, focuses on the key features of digital method for
App Masaru Honjo, app medicines for smartphone
Restriction Akio restriction smartphone addiction.
Function for Yoneyama, function. addiction, and analyze the
Smartphone Masato Uchida usage of the function and
Addiction verify its effectiveness.
2. Their results
showed significant
differences in both
psychological and
behavioral aspects
between those who used
the app restriction
function and those who
did not. Specifically, they
found that the app
restriction function is
more likely to be used by
those who were more
aware of their
smartphone addiction.

9 Human Iftakhar Here they 1. They have found The paper only
Behaviour Impact Mohammad have used out the major problem of focuses on the
to Use Talha, Imrus Naïve Bayes the human behaviour’s accuracy of the
of Salehin, theorem to negative side and its system and does not
Smartphones Susanta learn impact different sources like tends to give an
with the Chandra Debnath, of mental imbalance, stress, specific output to
Python Mohd. hum depression, loneliness, human behaviour
Implementatio n Saifuzzaman, an behaviour. etc. with respect to
Using Naive Nazmun 2. With the help of smartphone.
Bayesian naïve bayes theorem and
classifier, support vector
machine, special data set
of human behavior, and
probability are used to
calculate accuracy.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 6
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue III Mar 2022- Available at www.ijraset.com

10 Human Kostas This is also 1. This paper The tends

Behaviour Konsolakis, actually a surveys the to only
Analysis through Hermie survey state-of-the-art in understand human
Smartphones Hermens, paper which human behaviour behaviour
Claudia is analysis based on and
Villalonga, addressing smartphones. doesn’t tends to
Miriam the 2. They have use this analysis
VollenbroekHutten huma categorize prior for Furter
n behaviour works into four main prediction
analysis. sensing modalities proposed or else.
related to physical,
cognitive, emotional
and social behaviour
which will help the
smartphone user to
improve themselves.

III. CONCLUSION
A. The task of manually performing every search, every navigations involves a quit amount of time and can be also frustrating
task.
B. So this study shows us that a simple suggestions that follows the predictions done by the AI can make a difference in users
smartphone experience.

REFERENCES
[1] Jain, Ankita Kanhangad, Vivek [Supervisor], “Human behavior analysis• using smartphone sensor data”, Discipline of Electrical Engineering, IIT Indore,
8Mar-2019.
[2] Subrata Tikadar, Samit Bhattacharya, “How Do They Use Their• Smartphones: A Study on Smartphone Usage by Indian Students”, IIT Guwahati - Indian
Institute of Technology Guwahati, 2020.
[3] Hannu Verkasalo, “Analysis of Smartphone User Behavior”, IEEE, 2010
[4] Ruxia Hong, “Research on Mobile User Behavior Mining Model Based on Big Data”, IEEE, 2020.
[5] S. Sareh Ahmadi, Sherif Rashad, Heba Elgazzar, “Machine Learning Models for Activity Recognition and Authentication of Smartphone Users”, IEEE, 2020.
[6] Balaji Balasubramanian, Pranshu Diwan, Rajeshwar Nadar, Anuradha Bhatia, “Analysis of Facial Emotion Recognition” IEEE, 2019.
[7] Mahnaz Roshanaei, Richard Han, Shivakant Mishra, “EmotionSensing: Predicting Mobile User Emotions”, IEEE, 2017.
[8] Natasha Jaques, Sara Taylor, Asaph Azaria, Asma Ghandeharioun, Akane Sano, “Predicting students' happiness from physiology, phone, mobility, and
behavioral data”, IEEE, 2015.
[9] Jiqiang Song, Eugene Y. Tang, Leibo Liu, “User Behavior Pattern Analysis and Prediction Based on Mobile Phone Sensors”, Intel Labs China, Beijing, China.
[10] Gabriella M. Harari, “Using Smartphones to Collect Behavioral Data in Psychological Science: Opportunities, Practical Considerations, and Challenges”,
Department of Psychology, The University of Texas at Austin. Smartphone Addiction - Lawrence Robinson, Melinda Smith, M.A., and Jeanne Segal, Ph.D.
Teens and social media use: What's the impact? - Mayo Clinic Staff , How to Break a Bad Habit and Replace It With a Good One – JAMES CLEAR.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 7