Multimedia based IoT-centric smart framework for eLearning paradigm Muhammad Munwar & others

Multimedia based IoT-centric smart framework
for eLearning paradigm
Muhammad Munwar Iqbal1,2
& Muhammad Farhan3
&
Sohail Jabbar4
& Yasir Saleem2
& Shehzad Khalid5
Received: 24 July 2017 /Revised: 9 December 2017 /Accepted: 8 January 2018
# Springer Science+Business Media, LLC, part of Springer Nature 2018
Abstract Multimedia content boosts the learning trends. This paper is aimed to presents
an electronic learning system based on Internet of Things (IoT) for the synchronous and
asynchronous communications. The infrastructure of IoT provides the adaptable, scalable
and open access for the eLearning paradigm. The multimedia-based IoT-centric environ-
ment is suitable to enhance the effectiveness of the delivery of learning contents.
Students can take full advantage of 7As of IoT, which provides the opportunity to the
students that they can access everything on the internet at any time and place. It creates a
flexible eLearning paradigm for the teachers and students. The proposed eLearning
modeluses sensors to detect the student location, temperature, and mobile camera to
identify the student activeness in thelearning environment. Virtual campuses are con-
trolled from a centralized location that may be called the head office. The MAQAS
framework provides the solutions to the problems and analyzes the results for the
efficient and connected eLearning paradigm. The MAQAS system is used to answer
student’s queries, which are responded to automatically by agent-based question answer-
ing system. The results show that the students’ participation towards learning and
teacher’s pedagogy are more efficient in synchronous and asynchronous modes. Perfor-
mance evaluated by comparison to the existing question answering Live QA Trak, Quora
Yoda QA Live and AskMSR-QA with MAQAS.
Multimed Tools Appl
https://doi.org/10.1007/s11042-018-5636-y
* Sohail Jabbar
sjabbar.research@gmail.com
1
Department of Computer Science, University of Engineering and Technology, Taxila, Pakistan
2
Department of Computer Science and Engineering, University of Engineering and Technology,
Lahore, Pakistan
3
Department of Computer Science, COMSATS Institute of Information Technology, Sahiwal, Pakistan
4
Department of Computer Science, National Textile University, Faisalabad, Pakistan
5
Department of Computer Engineering, Bahria University, Islamabad, Pakistan
(2019) 78:3087–3106
/
Published online: 13 February 2018

Keywords Multimedia-centric Internet of Thing (mm-IoT) . FUNF. Question answering .
Multimedia contents . eLearning . Multimedia-aware IoTsystem . Sensors
1 Introduction
Electronic Learning (eLearning) is utilized to teach individuals nowadays. Various schools are
beginning distinctive courses by utilizing eLearning for secondary education level to degree
level and even at postgraduate level. This paper depicts the best-known distinctive machine
studying methods to help up the e-Learning training standard and model [2, 12]. Exhaustively
supervised and unsupervised systems are portrayed here for the e-Learning standard to the auto
answer of learners’ inquiries. Web bot is consolidated for the studying of people who are
taking courses in remote mode [15]. This paper focuses the multimedia-based IoT system
architectures and services used in eLearning paradigm. IoT is well defined, and it offers the
innovative techniques for the connectivity of devices with systems. The computing devices are
connected to existing infrastructure using IoT. IoT provides the variety of connection that can
be used for the Machine-To-Machine (M2M) communications. It also offers the connectivity
for the devices, system, and services by using different protocols. The smart object like smart
mobile phone, watches, and other personal handheld devices, as well as embedded devices, are
utilized for the automation of the all the fields of the life [10]. The IoT with multimedia offers
services multimedia data optimization, camera networks, complexity aware algorithms de-
signs, security algorithms, low power design, distributed network implementation and man-
agement, and for audio-video services [4, 8].
The mix of eLearning with IoT offers the utility computing as Software as a Services (SaaS).
The deployment of the eLearning resources and tools are accessible by using the connectivity
through IoT. IoT provides the services at less hardware requirement need and connects offers at
minimal cost. That leads to lessening the burden of the infrastructure cost. It allows the institution
to focus on their core business of teaching instead of concentrating on the search and purchase of
cost-effective module for the eLearning System [17]. It also helps them to obtain the latest
technology and updates without charges. IoT provides the following advantage in eLearning [8]:
& IoT delivers the better facilities for the students to get connected with their Institute and
teacher 24/7, even they are travelling from one place to another place
& A student may take the online exam by any of the internet connection through Wi-Fi,
GSM-based 2G, 3G, 4G and 5G.
& The IoT offers the connectivity to anyone, anytime and anywhere. The eLearning appli-
cation development and accessibility are demanding high web development skill.
& As student are connected through IoT, they have used to take back on OneDrive, Google
Drive, and another lot of space available by the other vendors.
& It also allows the students to work and take lectures from anyplace, i.e. library, home,
office, campus, workplaces.
& Student using the IoT having access to their files and data on the laptop, mobile and
desktop PC to use an update.
& Using IoT and internet-storage crash-recovery techniques are nearly unneeded.
IoT-based connection scenarios are too many that used to connect the devices for a short
time as well as permanent [1]. The connection modes used for the IoT device connectivity are
Multimed Tools Appl (2019) 78:3087–31063088

following: Wired technology, Wi-Fi technology, mobile GMS based connectivity, ZigBee,
Bluetooth, ISA100, 6LowPAN and Hi-Fi [27, 30, 38]. IoT connected scenario is shown in
Fig. 1.The student can connect the servers, printers, scanners, laptop, vehicles, tablet,
smartphones, healthcare units, IP phones, gadgets and sensors by the IoT-centric devices for
the collection of multimedia data. The eLearning needs connectivity with the learning man-
agement system and teacher on the priority basis for the student.
1.1 Contribution
The author contributed to this investigation that proposed a framework the Multimedia and
Agent-based Question Answering System (MAQAS). The MAQAS framework facilitates the
student and teacher interaction in an improved fashion. Multimedia base Internet of Things
provides the connectivity to the students and teachers with significant improvements. The
mechanism introduced in this study enhance the learning skill of the student. The student using
the IoT having location free access to their files and data on the laptop, mobile and desktop PC
to use an update. Moreover, the sensors used for the collection of data for the identification of
location, study and lecture area conditions and Mobi-came for the attentiveness of the students.
The manuscript is organized as follows: Section 2 is literature review which provides the
comparative and existing material about the multimedia, eLearning, and IoT. Section 3 is the
comprises of the proposed solution. In this section multimedia agent based question answering
system is discussed. The 4th Section gives the results and discusses the proposed solution and
comparisons with other question answering systems. References in 5th section proceed the
conclusion.
2 Literature review
The information values are increasing at high speed with a contribution in the interconnection
provided by the IoT for the information processing and transformations. IoT plays an excellent
Fig. 1 IoT Connected Scenario
Multimed Tools Appl (2019) 78:3087–3106 3089

role for the betterment of the society and humanity. IoT replaces the existing connectivity with
new trends and techniques. Now connectivity for the peoples becomes a service. People may
connect in seven A’s fashion. The IoT provides the connectivity for Anywhere, Anytime,
Anyplace, with Anyone, Anything using Any-network path for any service, which the students
are required. This well defined in the ITU vision which describes the four A’s, i.e. anyplace,
anytime, for anyone, and connectivity for anything [3, 22] as shown in Fig. 2.
2.1 Asynchronous communication
In eLearning paradigm, synchronous and asynchronous types of communication tools are used
by the student-teacher interactions [19]. Synchronous communication tools are used for the
group chat, one-to-one using Skype, TeamViewer, and so on. The devices utilized for the
synchronous communication have the fast data transfer rates. In asynchronous techniques, the
student and teacher are present in virtual space for communication. Asynchronous does not
need to send and receive the data in real time. It is slower than synchronous communications,
i.e. electronic mail, Moderated Discussion Boards. Student by using the asynchronous tools
repeat the lectures as many times as they required. The asynchronous tools provide the
facilities to students to participate in group studies. It enhances the learning of the student
by the group activities. The use of the asynchronous tools eliminates the preconceived notions
about the race, sex, and colour. These tools are time efficient and profitable.
2.2 The role of IoT in eLearning
Virtual campuses are controlled from a centralized location that may be called the head office.
The LMS may contain the portal software for course management [20]. LMS will have
additional tractability in the plan and incorporation of intelligent agents. The agent is a piece
of code with persistent ownership and control of working mechanism. In-house designed and
development of agents can be more skilful, more accessible and faster. However, this is only
feasible for more prominent institutes with better programming and database expertise, as well
as substantial resources [7, 19]. These institutions may have IT support units and cloud
IoT
Anytime
Anywhere
Any
Hardware
Anyone
Anything
Any
Network
Any-service
Fig. 2 IoT 7A’s Connectivity

services or more research groups within their academic departments [13, 21, 28].FUNF-
Journal allows the user or researcher to configure data collection parameters for over 30
different built-in data probes, including all phone sensors, as well as additional data types and
high-level probes that generate inferences and new data based on the output of sensor data.
Further probes are continuously added to the system. The app supports importing and
exporting of probing configurations, as well as remote configuration (with user permission).
The FUNF tool is based on the simple configurations that data for investigation is
automatically gathered if the smart mobile phone used in daily routine life. This data is stored
securely on the smart cell phone in an encrypted format. This information used to be extracted
from the smart cell phone in many ways. Data collected by manually exporting is done via
email or any other Android service. Data can be transferred by manually copying to another
device by the memory card. Data also be collected by linking DropBox account with
application configuring FUNF to auto uploading using IoT services. When the server config-
uration is set up, the data collection configuration could also be performed remotely. Once the
data is extracted from the phone, we provide a set of desktop utilities that allow decryption of
the data, demonstrate some examples of visualizing and considering the collected data.
2.2.1 Student-teacher interaction systems
The automated question answering (QA) track, which has been one of the most popular tracks in
TREC for many years, focuses on the task of providing automatic answers to human questions.
The Question Answering track primarily deals with factual questions. The responses of the
participants are extracted from the News article corpus. While the task evolves to model increas-
ingly practical information needs, addressing question series, list questions, and even interactive
feedback, a significant limitation remains: the questions do not directly come from real users in real
time. The LiveQA track revives and expands the QA track, focusing on Blive^ question answering
real-user questions this year [9]. Real user questions, extracted from the stream of most recent
questions submitted on the Yahoo Answers (YA) site that has not yet been answered by humans,
will be sent to our systems. Then our system provides an answer in real time [9].
Quora is a fast-growing social Q&A site where users create and answer questions, and
identify the best answers by up-votes and down-votes with crowd wisdom. The Web of Data
has grown to contain billions of facts about a broad variety of domains. While experts can
easily access this wealth of data, it remains difficult to use for non-experts [25]. The pre-
implemented HAWK system, however, performs better on hybrid questions than on plain
English questions. Systems like YODA, which do only provide answers without a SPARQL
query cannot be analyzed sufficiently [32]. AskMSR to a new question answering dataset
created using historical TREC-QA questions; we find that the critical step, query pattern
generation is no longer required but instead, more in-depth NLP analysis on questions and
snippets remains critical [31].
3 Proposed Multimedia and Agents based Question Answering System
(MAQAS)
The Multimedia and Agents based Question Answering System (MAQAS) in eLearning paradigm
help the student to find answering more relevant and accurate. The proposed solution is based on
multimedia data for the eLearning systems [18]. This solution is agent-based to facilitate the student

and utilization of the resources efficiently. Text Mining comprises of the pattern discovery from the
contextual documents. Mobile Sensor for Student data collection using FUNF is shown in Fig. 3.
The Multimedia-centric Internet of Things (mm-IoT) is the collection of interfaces, proto-
cols and associated multimedia-related information representations that enable advanced
services and applications based on human-to-device and device-to-device interaction in
physical and virtual environments. Information refers to data sensed and processed by a
device, and communicated to a human or another device [26]. The IoT-based for a smart
framework for the eLearning is used to student-student interaction as well as for the student
teacher interaction [16]. Student-teacher collaboration in IoT scenario is shown in Fig. 4.
The smart mobile phones are now integrated with biometric sensors which used with Closed-
Circuit Television (CCTV) camera to a highly sensitive location in the city [23, 34]. Sensors are
used to make the reality, and now sensors are shifting from smart to intelligent manner [11].
Biometric is also used for the for the identification of the individuals same as computerized national
identity card [35]. Billions of the sensors are deployed and connected through the smart mobile
phones with IoT, and much more are connected through networks protocols with these smart
mobile phones. The data is collected from the sensors are an interoperable format for creating and
developing systems that break the soil vertically and harvest in cross-domain that is useful for the
application connects to IoT. In eLearning paradigm, the IoT applications are used no most widely.
Multimedia-based Student’s Interaction with Intelligent [11, 37] LMS is shown in Fig. 5.
3.1 Student connectivity technologies
In eLearning paradigm, students have some connectivity methods used to connect with LMS
and teacher. The most widely used method is campus-based student uses the Wi-Fi internet
services available on campus. The other methods are used for the connectivity of the IoT
devices possessed by the students as shown in Fig. 6.
The mobile internet, which is mostly used by the countryside student where internet line is
not available. The mobile phone internet connection technologies are categorized into four
types: 2G, 3G, 4G and 5G as shown in Table 1. The mobile internet services are provided by
the different companies PTCL, Mobilink, Warid, Ufone, Zong, and Telenor [14, 24].
Notification
Applications
Reports & Studies
You DB Choice
Visulization
Location
Movement
Man More and
3rd Party APIs
Usage/Communication
Social Proximity
Open Sourcing
Framework
Fig. 3 Mobile Sensor for Student data collection using FUNF (http://inabox.funf.org)

Wireless Technolog
for Connectivity
Network as a
Service/
Networking
Services
Mobile
Sensors
Multimedia Content
Delivery Network
Multi-Screen Technology
for Content Sharing
OTT
Machine to
Machine
Connections
eLearning
Content
Security
Management
FFTx
Zigbee
Connection
Unified and Non-Unified
Communication and
Collaboration
Corporate Internet Connect
Wi-Fi
2G, 3G,
4G
2G
3G
4G
Wi-Fi
Teacher Delivering Video Lecture
PTCL Broadband
PTCL Broadband
PTCL Broadband
Embedded Technology
Networking
Wi-Fi
Fig. 4 Student Teacher Collaboration in different IoT Connectivity Standards in MAQAS
Teacher
Student 2Mobile Sensors
TEPM
GPS
MobiCame
Mobile Sensors
TEPM
GPS
MobiCame
Student 2
Mobile Sensors
TEPM
GPS
MobiCame
Student 3
Mobile Sensors
TEPM
GPS
MobiCame
Student 4
Mobile Sens
TEPM
GPS
MobiCame
Student N
Lectures : Youtube.com
VUTV 1, 2, 3, 4 & CDs
Handouts: PDF Files,
Word Docs, etc.
Lecture Slides: Power
Point, PDF Files, etc.
Books: PDF Files, DJVU
Files, etc.
LMS: Student
LMS: Moderated
Discussion Board
LMS: Admin
SMS Services
Multimedia Based Intelligent LMS
Fig. 5 Multimedia and IoT based Infrastructure for Student’s Interaction with Intelligent LMS

The coverage of the companies is dependent upon directly at PTCL services. PTCL
provides the backbone infrastructure for the telecommunications services. These internet
service provider companies located Karachi to Skardu in Pakistan. Students use these services
to connect LMS in eLearning paradigm. Teacher commonly used the corporate internet service
provided by the Higher Education Commission at offices. The teacher uses the higher
bandwidth as interacted with some students for their problem-solving.
3.2 Data collection from IoT-based distributed sensors
Data collection from IoT-based distributed sensors is done with the help of following modules.
& Mob-Cam for Student Attentiveness
Mobile camera is used to measure the student attentiveness by the student face recognition.
The studentis either taking the video lecture or talking to some other student [6, 36].
Fig. 6 Classification of IoT Modules w.r.t Energy, Capacities and Micro-batteries
Table 1 Mobile Services for Student Connectivity
Companies Postpaid Prepaid 2G 3G 4G
PTCL ✓ ✓ ✓ ✓ ✓
Mobilink ✓ ✓ ✓ ✓ ✘
Warid ✓ ✓ ✓ ✓ ✓
Ufone ✓ ✓ ✓ ✓ ✘
Zong ✓ ✓ ✓ ✓ ✓
Telenor ✓ ✓ ✓ ✓ ✘

& Temperature Data
The temperature of the location is detected by using the mobile sensors for the analytical
purposes. This dataset is used for the further recognition of the conditions in which student are
listening video lectures [6].
& Location Data
The student coordinates are collected through smartphone GPS sensor as shown in Table 2.
These coordinates are helpful to detecting the location of the student. Once the student location
is identified then decides for either immediate appointment is possible in case teacher has a
free slot. Otherwise, appointment deferred to some appropriate time as shown in Table 2.
Appropriate time will have sought out by a collaboration of the student. While also
communicate to the teacher for the available free slot so, that student teacher synchronous
interaction is possible for a better solution of the student problem. Now Google Map is used to
find out the locations of students. User smartphones are used to collect their Global Positioning
System (GPS) coordinates. Google Maps API has used the location into coordinates like
latitude and longitude dynamically, and these coordinates are used to place markers on the map
as shown in Fig. 7. Multiple markers show locations of students, whether they are on the
campus or somewhere else while using the framework which is based on multimedia in
eLearning as shown in Fig. 7.
Now agent A2 collects the feedback from the student. This feedback along with question
relevancy used to measure the need for multimedia-based communications. If the student
Table 2 A Chunk of Student’s Data Collected using Smartphone GPS
Student Number Latitude Longitude Location City Sea Level
1 31.6533814 74.29504395 Home Lahore 216 m
2 31.5949131 74.30294037 Office Lahore 219 m
3 31.5779505 74.31598663 Campus Lahore 223 m
4 31.5691755 74.31049347 Workplace Lahore 216 m
5 31.5685904 74.32971954 On-road Lahore 219 m
6 31.5644951 74.30774689 Office Lahore 217 m
7 31.5838010 74.33933258 Campus Lahore 219 m
8 31.5568891 74.30706024 Workplace Lahore 213 m
9 31.5018730 74.31049347 Office Lahore 218 m
10 32.1593377 74.17762756 Campus Gujranwala 216 m
11 32.1471297 74.20131683 Workplace Gujranwala 227 m
12 32.1561405 74.20183182 On-road Gujranwala 230 m
13 32.1497458 74.19565201 Office Gujranwala 223 m
14 32.1514899 74.17676926 Campus Gujranwala 227 m
15 32.5097617 74.55596924 Workplace Sialkot 251 m
16 32.4842801 74.51614380 Office Sialkot 250 m
17 32.5792206 74.08218384 Campus Gujrat 237 m
18 33.6020382 73.05839539 Office Rawalpindi 498 m
19 33.6020382 73.04191589 Office Rawalpindi 502 m
20 33.6980652 73.06114197 Campus Islamabad 546 m
21 33.7014928 73.08860779 Workplace Islamabad 533 m
22 33.7026353 73.04191589 Office Islamabad 550 m
25 33.6603531 73.06800842 Office Islamabad 512 m

feedback is satisfactory, then the agent A2 will fix major role-plays in the next step towards
appointment. In asynchronous mode, students have, limited facility to post their queries
through LMS based interface.
These queries are commonly text-based, and no multimedia facility is available. In
MAQAS proposed model, student’s queries are answered automatically by agent-based
question answering system. The teacher answers selected queries locked by the system.
Teacher reply is also containing the excessive text with graphics are optional. Therefore,
there is a strict need to introduce a framework for the student-teacher interaction that
provides the multimedia-based communication to satisfy the student queries. IoT module
classification regarding power management, energy storage [29], radio frequency, sens-
ing modules is shown in Fig. 6.That is based on instructor feedback, which is being
validated by the MAQAS and suitable for both distances learning and the asynchronous
online environment by merging this solution into existing management scenario. This
scenario prevents students from unnecessary tension in getting their reply in a scenario in
which teacher is not physically present [5, 33]. To shorten incredible flow of research in
Fig. 7 A Subset of the data of student cluster on Google Map

the field of signals, audio, video, sensors and imaging informatics is ongoing. It may
result in the determination process for this area Bsignals, audio, video, sensors and
imaging informatics" which carried out brilliant articles, speaking to research in four
separate countries. The fields of cerebrum machine interfaces, sound observation in
telemonitoring, delicate tissue exposing, and body sensors have been chosen. The
segment can just mirror a little parcel of the overall productive work in the field of
signals, audio, video, sensors and picture transforming with applications in stimulating
informatics.
4 Results and discussion
Let’s run a simple query Bwhat is the role of the operating system?^ on MAQAS system and
take a subset of the keywords and calculate the relevance for the decision-making at a later
stage to fix appointments of the student-teacher interaction in synchronous communications.
Table 2 denoted the keywords relevancy on small subset keywords of the answer delivered to
the student that is calculated. Now agent A2 collects the feedback from the student. This
feedback along with question relevancy used to measure the need for multimedia-based
communications. If the student feedback is satisfactory, then the agent A2 will fix major
role-plays in the next step towards appointment. Agent A2 collects the coordinates through
smartphone GPS sensor as shown in Table 2. These coordinates are helpful to detecting the
location of the student. Once the student location is identified, then agent A2decides for either
immediate appointment is possible in case teacher has a free slot. Otherwise, appointment
deferred to some appropriate time.
4.1 Comparison of student teacher interaction systems
Comparison of MAQAS system with other question answering systems for definition
type questions of the WH-type category. The accuracy of the multimedia and agent-based
question answering system is compared to the analysis of WH-type definition questions
for the accuracy of answers. It is analyzed for the 150 question which is comprised of the
definition type answer required for the question. The accuracy of our proposed system
MAQAS is 94%, and another system like LIVE QA TRAK for the same queries is 92%,
QUORA with 68%, YODA QA LIVE with 82% and ASKMSR-QA have the 87%
accuracy as shown in Fig. 8. The solid black line represents the polynomial accuracy
trend line for the total correct answers for WH definition type. The yellow dotted line
shows the linear accuracy trend line for the WH definition type complete correct
answers.
The accuracy of the multimedia and agent-based question answering system is com-
pared to the analysis of WH-type descriptive questions for the accuracy of answers. It is
analyzed for the 150 question which is comprised of the description type answer required
for the question. The accuracy of our proposed system MAQAS is 92%, and another
system like LIVE QA TRAK for the same queries is 89.33%, QUORA with 67%, YODA
QA LIVE with 74.67% and ASKMSR-QA have the 87.33% accuracy as shown in Fig. 9.
The solid black line represents the polynomial accuracy trend line for the total correct
answers for the Non-WH descriptive type. The yellow dotted line shows the linear
accuracy trend line for the Non-WH descriptive type complete, correct answers.

The accuracy of the multimedia and agent-based question answering system is
compared to the analysis of WH-type factoid questions for the accuracy of answers.
It is analyzed for the 150 question which is comprised of the factoid type answers
required for the question. The accuracy of our proposed system MAQAS is 95.33%,
and another system like LIVE QA TRAK for the same queries is 92%, QUORA with
80%, YODA QA LIVE with 94%, and ASKMSR-QA has the 86% accuracy as shown
in Fig. 10. The solid black line represents the polynomial accuracy trend line for the
150 150 150 150 150
142
138
102
123
131
94.67 92
68
82 87.33
0
20
40
60
80
100
120
140
160
MAQAS SYSTEM LIVE QA TRACK QUORA YODA QA LIVE ASKMSR-QA
ycaruccAdnasnoitseuQlatoT
Diﬀerent Systems for InteracƟon
Total QuesƟons Total Correct Answers Accuracy (%age)
Poly. (Accuracy (%age)) Linear (Accuracy (%age))
Fig. 8 Comparative analysis of WH-type definition questions for accuracy of answers
150 150 150 150 150
138 134
101
112
131
92
89.33
67.33
74.67
87.33
0
20
40
60
80
100
120
140
160
Fig. 9 Comparative analysis of WH-type descriptive questions for accuracy of answers

total correct answers for Non-WH factoid type. The yellow dotted line shows the linear
accuracy trend line for the Non-WH factoid type total correct answers.
The accuracy of the multimedia and agent-based question answering system is com-
pared to the analysis of Non-WH type definition questions for the accuracy of answers. It
is analyzed for the 150 question which is comprised of the definition type answers required
for the question. The accuracy of our proposed system MAQAS is 92%, and another
system like LIVE QA TRAK for the same queries is 90.67%, QUORA with 75.33%,
YODA QA LIVE with 78.67%, and ASKMSR-QA has the 91.33% accuracy as shown in
Fig. 11.The solid black line represents the polynomial accuracy trend line for the total
correct answers for Non-WH definition type. The yellow dotted line shows the linear
accuracy trend line for the Non-WH definition type total correct answers.
compared to the analysis of Non-WH type descriptive questions for the accuracy of
answers. It is analyzed for the 150 question which is comprised of the descriptive type
answers required for the question. The accuracy of our proposed system MAQAS is
92.67%, and another system like LIVE QA TRAK for the same queries is 88%, QUORA
with 80%, YODA QA LIVE with 69.33%, and ASKMSR-QA has the 90.67% accuracy
as shown in Fig. 12. The solid black line represents the polynomial accuracy trend line
for the total correct answers for the Non-WH descriptive type. The yellow dotted line
shows the linear accuracy trend line for the Non-WH descriptive type complete and
correct answers.
compared to the analysis of Non-WH type MCQs questions for the accuracy of answers.
It is analyzed for the 150 question which is comprised of the MCQs type answers
required for the question. The accuracy of our proposed system MAQAS is 96%, and
another system like LIVE QA TRAK for the same queries is 90.67%, QUORA with 38%,
YODA QA LIVE with 65.33% and ASKMSR-QA have the 87.33% accuracy, as shown
150 150 150 150 150
143
138
120
141
129
95.33 92
80
94
86
0
20
40
60
80
100
120
140
160
Fig. 10 Comparative analysis of WH-type factoid questions for accuracy of answers

in Fig. 13. The solid black line represents the polynomial accuracy trend line for the total
correct answers. The yellow dotted line shows the linear accuracy trend line for the total
correct answers.
The accuracy of the multimedia and agent-based question answering system is calculated
for the analysis of WH-type descriptive, factoid and definition questions for the accuracy of
150 150 150 150 150
138 136
113
118
137
92
90.67
75.33 78.67
91.33
0
20
40
60
80
100
120
140
160
Fig. 11 Comparison of MAQAS System with other Question Answering Systems for Definition Type Questions
of Non-WH Type Category
150 150 150 150 150
139
132
120
104
136
92.67 88
80
69.33
90.67
0
20
40
60
80
100
120
140
160
Fig. 12 Comparison of MAQAS System with other Question Answering Systems for Descriptive Type
Questions of Non-WH Type Category

answers. It is analyzed for the 150 question which is comprised of the descriptive, factoid and
definition type answers required for the question.
The overall accuracy of our proposed system MAQAS for the definition type is 94.67%, for
the descriptive type is 92%, and for the factoid, the type is 95.33% as shown in Fig. 14. The
dotted black line represents the polynomial accuracy trend line for the percent accuracy. The
orange dotted line shows the linear accuracy trend line.
150 150 150 150 150
144
136
57
98
131
96
90.67
38
65.33
87.33
0
20
40
60
80
100
120
140
160
180
Fig. 13 Comparison of MAQAS System with other Question Answering Systems for MCQ’s Type Questions of
Non-WH Type Category
94.67
92
95.33
88
89
90
91
92
93
94
95
96
97
DeﬁniƟon Type DescripƟve Type Factoid Type
PercentAccuracy
QuesƟon Types
MAQAS System Poly. (MAQAS System) Linear (MAQAS System)
Fig. 14 Overall Performance of MAQAS System for WH-Type Questions

5 Conclusion
The outcomes of this research are the design of the intelligent framework with the help of
advanced tools and technologies, which are used more efficiently to facilitate the student.
Question answering in the eLearning paradigm plays a crucial role to enhance the student
learning. Student learning process in affected severely if the answering is delayed. The
eLearning solution can be developed with the current technologies that lead to the multi-
media data. The analytics presents the learning behaviour, outcomes, and the learning
targets. The analysis is significant by the study of the relationship between multimedia data
used for the academic and student-teacher interactions for educational performance eval-
uation and enhancements. This framework supports the main eLearning components for the
convenience, easiness, accessibility, connectivity, and flexibility. eLearning system is
mostly facing challenges of optimizing resource that can be handled by the multimedia
and IoT-centric devices. Multimedia contents are fully utilized for the enhancement of the
learning curve of the students using IoT based infrastructure. eLearning with multimedia
settled the goodness of an IoT-based solution. The advantages focusing the integration of an
eLearning system into the IoT-based system can be highlighted as good flexibility and
scalability for the resources, including storage, computational with internet connectivity.
However, these are just initial steps towards an open line for research and exploitation of
eLearning and IoT-based computing platforms. In future, the proposed model will be used
in healthcare, engineering with more internet-connected devices. The increase in the
number and types of the device open the new horizon for the students and learners.
References
1. Baccelli E, Mehlis C, Hahm O, Schmidt TC, Wählisch M (2014) Information centric networking in the IoT:
experiments with NDN in the wild. arXiv preprint arXiv:1406.6608
2. Dolog P, Henze N, Nejdl W, Sintek M (2004) Personalization in distributed e-learning environments. In:
Proceedings of the 13th international World Wide Web conference on Alternate track papers & posters, pp
170–179
3. Domingo MC (2012) An overview of the Internet of Things for people with disabilities. J Netw Comput
Appl 35:584–596
4. Effelsberg W (2013) A personal look back at twenty years of research in multimedia content analysis. ACM
Trans Multimed Comput Commun Appl 9(1s):43
5. Farhan M, Iqbal MM, Aslam M, Enriquez AM, Farooq A, Tanveer S et al (2012) Automated Reply to
Students' Queries in E-Learning Environment Using Web-BOT. In: Artificial Intelligence (MICAI), 2012
11th Mexican International Conference on, pp 63–65
6. Farhan M, Iqbal MM, Naeem MR (2015) Analysis of Video Lecture's Images for the Compression Formats
Support for the Students in eLearning Paradigm. Sci Int 27:1171–1176
7. Farhan M, Aslam M, Jabbar S, Khalid S (2016) Multimedia based qualitative assessment methodology in
eLearning: student teacher engagement analysis. Multimed Tools Appl 77:4909–4923 https://doi.
org/10.1007/s11042-016-4212-6
8. Grois D, Hadar O (2013) Recent trends in online mutimedia education for heterogeneous enduser devices
based on Scalable Video Coding. In 2013 IEEE Global Engineering Education Conference (EDUCON),
IEEE, pp. 1141–1146
9. Höffner K, Walter S, Marx E, Usbeck R, Lehmann J, Ngonga Ngomo AC (2017) Survey on challenges of
question answering in the semantic web. Semantic Web 8(6):895–920
10. Iqbal MM, Farhan M, Saleem Y, Aslam M (2014) Automated web-bot implementation using machine
learning techniques in eLearning paradigm. J Appl Environ Biol Sci 4:9
11. Iqbal MM, Muhammad F, Yasir S, Muhammad A (2014) Automated Web-Bot Implementation using
Machine Learning Techniques in eLearning Paradigm. J App Environ Biol Sci 4:90–98

12. Iqbal MM, Farhan M, Saleem Y (2015) Personalization with Agent in E-Learning Framework. Sci Int 26:
2541–2543
13. Jabbar S, Naseer K, Gohar M, Rho S, Chang H (2016) Trust model at service layer of cloud computing for
educational institutes. J Supercomput 72:58–83
14. Jacobsen WC, Forste R (2011) The wired generation: academic and social outcomes of electronic media use
among university students. Cyberpsychol Behav Soc Netw 14(5):275–280
15. Juwattana P, Prompoon N (2012) Implementing Software Requirements and Modeling Using Mapping
Rules Based on B-SCP Framework Analysis. Int J Comput Commun Eng 1:135
16. Khalid S, Sajjad S, Jabbar S, Chang H (2017) Accurate and efficient shape matching approach using
vocabularies of multi-feature space representations. J Real-Time Image Proc 13(3):449–465
17. Khan MA, Iqbal MM, Ubaid F, Amin R, Ismail A (2016) Scalable and Secure Network Storage in Cloud
Computing. Int J Comput Sci Inf Secur 14:545
18. Li Z, Huang H, Misra S (2017) Compressed sensing via dictionary learning and approximate message
passing for multimedia internet of things. IEEE Internet Things J 4:505–512
19. Malik KR, Ahmad T, Farhan M, Aslam M, Jabbar S, Khalid S, Kim M (2016) Big-data: transformation
from heterogeneous data to semantically-enriched simplified data. Multimed Tools Appl 75(20):12727–
12747
20. Munwar Iqbal MFM, Saleem Y, Aslam M (2014) Automated Web-Bot Implementation using Machine
Learning Techniques in eLearning Paradigm. J Appl Environ Biol Sci 4:9
21. Naseer MK, Jabbar S, Zafar I (2014) A novel trust model for selection of Cloud Service Provider. In:
Computer Applications & Research (WSCAR), 2014 World Symposium on, pp 1–6
22. Osseiran A, Boccardi F, Braun V, Kusume K, Marsch P, Maternia M et al (2014) Scenarios for
5G mobile and wireless communications: the vision of the METIS project. IEEE Commun Mag
52:26–35
23. Ozcan A (2014) Mobile phones democratize and cultivate next-generation imaging, diagnostics and
measurement tools. Lab Chip 14:3187–3194
24. Palwasha RI (2016) Using BBeyond the Quick Fix model^ identify the challenges of PTCL and suggest as
OD consultant for the development of this organization. Int J Exp Learn Case Stud 1:17–33
25. Patil S, Lee K (2016) Detecting experts on Quora: by their activity, quality of answers, linguistic
characteristics and temporal behaviors. Soc Netw Anal Min 6(1):5
26. Poluru RK, Naseera S (2017) A literature review on routing strategy in the internet of things. J Eng Sci
Technol Rev 10(5):50–60
27. Ramlee RA, Leong MH, Sarban Singh RSA, Ismail MM, Othman MA, Sulaiman HA et al. (2013)
Bluetooth remote home automation system using android application :1–5
28. Razzaq Malik K, Umar Chaudhry M, Munwar Iqbal M, Saleem Y, Farhan M (2014) Data Security and
Privacy in Cloud Computing: Threat Level Indications. Sci Int 26:1991–1996
29. Razzaq Malik K, Ahmad T, Farhan M, Ullah F, Amjad K, Khalid S (2016) Multiagent semantical
annotation enhancement model for iot-based energy-aware data. Int J Distrib Sens Netw 12(6):9103265
30. Shariff F, Rahim NA, Hew WP (2015) Zigbee-based data acquisition system for online monitoring of grid-
connected photovoltaic system. Expert Syst Appl 42:1730–1742
31. Sun H, Ma H, Yih WT, Tsai CT, Liu J, Chang MW (2015) Open domain question answering via semantic
enrichment. In Proceedings of the 24th International Conference on World Wide Web. International World
Wide Web Conferences Steering Committee, pp. 1045–1055
32. Usbeck R, Ngomo ACN, Haarmann B, Krithara A, Röder M, Napolitano G (2017) 7th Open Challenge on
Question Answering over Linked Data (QALD-7). In Semantic Web Evaluation Challenge. Springer, Cham,
pp. 59–69
33. Wang T-H (2014) Developing an assessment-centered e-Learning system for improving student learning
effectiveness. Comput Educ 73:189–203
34. Yang J, He S, Lin Y, Lv Z (2017) Multimedia cloud transmission and storage system based on internet of
things. MultimedTools Appl 76:17735–17750
35. Yuan J, Yu S (2013) Efficient privacy-preserving biometric identification in cloud computing. In:
INFOCOM, 2013 Proceedings IEEE, pp 2652–2660
36. Zain-Ul-Abidin, Farhan M, Iqbal MM, Naeem MR (2015) Analysis of Video Lecture's Images for the
Compression Formats Support for the Students in eLearning Paradigm. Sci Int 27:1171–1176
37. Zhang B, Han J, Shao L (2017) Guest Editorial: Feature Learning from RGB-D Data for Multimedia
Applications. MultimedTools Appl 76:4243–4248
38. Zhu Q, Wang R, Chen Q, Liu Y, Qin W (2010) Iot gateway: bridging wireless sensor networks into internet
of things. In 2010 IEEE/IFIP 8th International Conference on Embedded and Ubiquitous Computing
(EUC), IEEE, pp. 347–352

Dr. Muhammad Munwar Iqbal is serving as Assistant Professor in Computer Science Department at
University of Engineering and Technology, Taxila, Pakistan from more than three years. As an assistant professor
along with teaching activities bearer following responsibilities: Director Academic Cell, Head of the Semester
Committee, Head of the Scholarship Committee, HEC Laptop Scheme (Focal Person), Security Focal Person,
Head of the Exam Scrutiny Committee, Curriculum Revision Committee, Advisor COMPTECH Society,
Advisor BSCS 2016-17 Session and BS Cyber Technology 2017 and lead Prospectus Amendment Committee.
I have served in Computer Science department of Virtual University of Pakistan more than nine years. I have
received my Doctoral of Philosophy in Computer Science degree from University of Engineering and Technol-
ogy, Lahore, Pakistan. I have completed my MS in computer science from COMSATS Institute of Information
Technology, Lahore. I have completed my Master in Computer Science for the University of the Punjab Lahore. I
have first class education career throughout 20 years of the academia. More than 12 years of professional career
in renowned public-sector organizations. I have published 40 research papers in impact factor journal, conference
and local journals including Springer, Elsevier, TextRoad, Modestum and Hindawi, and conference proceedings
of IEEE. I have high practical implementation skills in Professional as well as academics Perspective. I have high
analytical and problem-solving abilities. My research interests includes Data Warehousing, Data Mining, e-
Learning, Machine Learning, Web Semantics, Artificial Intelligence and Data Science.
Muhammad Farhan is Assistant Professor at COMSATS Institute of Information Technology (CIIT),
Sahiwal Campus, Pakistan. He was working as lecturer in CIIT, Sahiwal. He has been working as

instructor of computer sciences in Virtual University of Pakistan (VU). Before joining to the university
he was also working as lecturer in a college. He is has completed his PhD from Department of
Computer Sciences and Engineering in University of Engineering and Technology (UET), Pakistan. He
obtained MSCS from University of Management and Technology (UMT), Pakistan. He has received
BSCS from Virtual University of Pakistan (VU). Currently he has 11+ years of teaching experience.
He has supervised more than 50 undergraduate projects. His research work is published in various
renowned journals of Springer, Elsevier, TextRoad, Modestum and Hindawi, and conference proceed-
ings of IEEE. He has been the reviewer for leading journals (MTAP, FGCS, CHB, TJS, and EIJ,
among many) and conferences (C-CODE 2017, ACM SAC 2016, WOSTECH17, WECNET17, among
others). He is currently engaged as TPC member chair in many conferences. He is a SIGAPP ACM
member. His research interests include, e-Learning, Computer Programming, Machine Learning,
Computer Vision, and Databases.
Dr. Sohail Jabbar is an Assistant Professor at National Textile University, Faisalabad Pakistan. He was
Post-Doctoral Researcher with Kyungpook National University, Daegu, South Korea. He also served
as Assistant Professor in the Department of Computer Science, COMSATS Institute of Information
Technology (CIIT), Sahiwal and also headed Networks and Communication Research Group at CIIT,
Sahiwal. He received many awards and honors from Higher Education Commission of Pakistan,
Bahria University, CIIT, and the Korean Government. Among those awards, Best Student Research
Awards of the Year, Research Productivity Award, BK-21 Plus Post Doc. Fellowship are few. He
received the Research Productivity Award from CIIT in 2014 and 2015. He has been engaged in
many National and International Level Projects. He has authored 1 Book, 2 Book Chapters and 60+
research papers. His research work is published in various renowned journals and magazines of IEEE,
Springer, Elsevier, MDPI, Old City Publication and Hindawi, and conference proceedings of IEEE,
ACM, and IAENG. He has been the reviewer for leading journals (ACM TOSN, JoS, MTAP,
AHSWN, ATECS, among many) and conferences (C-CODE 2017, ACM SAC 2016, ICACT 2016,
among others). He is currently engaged as TPC memberchair in many conferences. He is guest editor
of Sis in Future Generation Computer Systems (Elsevier), Peer-to-Peer Networking and Applications
(Springer), Journal of Information and Processing System (KIPS), and Cyber-Physical System (Taylor
& Francis). Sohail is on collaborative research with renowned research centers and institutes around
the globe on various issues in the domains of Internet of Things, Wireless Sensor Networks and Big
Data.

Dr. Yasir Saleem is currently serving as an Associate Professor in University of Engineering and Technology (UET),
Lahore, Pakistan. His research interests include computer networks, power electronics, digital signal processing and
control system. He completed his secondary education (O-level and A-level) form England. He had achieved his
Bachelor, Master and Ph.D. degree from Electrical Engineering Department of UET in 2002, 2004 and 2011
respectively. During his Ph.D. he has worked in Energy Conversion Lab, Universiti Technologi Malaysia (UTM) for
one semester under supervision of Prof. Dr. Zainal Salam who is Professor in Power Electronics and Renewable Energy,
Faculty of Electrical Engineering, Universiti Teknologi Malaysia, Malaysia. He has supervised 10+ M.Sc. theses and 05
PhD scholars are enrolled under his supervision. Currently he has authored / co-authored 20+ journal publications and
10+ conference papers at national and international level in field of Computer / IT and Electrical Engineering.
Dr. Shehzad Khalid is a Professor at Department of Computer Engineering, and Director of Post Graduate
Programs, Bahria University, Pakistan. He is a qualified academician and researcher with more than 60 international
publications in various renowned journals and conference proceedings. Dr. Shehzad has graduated from Ghulam
Ishaq Khan Institute of Engineering Sciences and Technology, Pakistan, in 2000. He received the M. Sc. degree from
National University of Science and Technology, Pakistan in 2003 and the Ph.D. degree from the University of
Manchester, U.K., in 2009. Dr. Shehzad is the Head of Computer Vision and Pattern Recognition (CVPR) research
group which is a vibrant research group undertaking various funded research projects. His areas of research include
but are not limited to Shape analysis and recognition, .Motion based data mining and behavior recognition, Medical
Image Analysis, ECG analysis for disease detection, Biometrics using fingerprints, vessels patterns of hands/retina of
eyes, ECG, Urdu stemmer development, Short and Long multi-lingual text mining, Urdu OCR etc. Dr. Shehzad has
been the reviewer for various leading ISI indexed journals. He has received Best Researcher Award for the year 2014
from Bahria University. He has also been awarded Letter of Appreciation for Outstanding research contribution in
year 2013 and outstanding performance award for the academic year 2013-2014.

Multimedia based IoT-centric smart framework for eLearning paradigm Muhammad Munwar & others

More Related Content

What's hot (20)

Similar to Multimedia based IoT-centric smart framework for eLearning paradigm Muhammad Munwar & others (20)

More from eraser Juan José Calderón (20)

Recently uploaded (20)

Multimedia based IoT-centric smart framework for eLearning paradigm Muhammad Munwar & others