Determine the User satisfaction on E-learning using the...

Determine the User satisfaction on E-learning using

the Decision Tree Algorithm with Kano Analysis

Z. Sheik Kamaludeen #1

, Dr. V. Uma Devi *2

, #1*2

Department of Computer Science , Jairams Arts and Science College, Karur, Tamilnadu, India.

[email protected] [email protected]

Abstract— The aim of this research work is to study the different types of e-learning and understands its interface and also determines

the learner’s satisfaction on using the e-learning method. The main purpose of this research work is by applying the data mining

technique such as, FAM-CART algorithm with Decision Tree algorithm to find out the anomalies and using the Kano analysis to

determine the learners’ Satisfaction on using the E-Learning. This work will be focused on learner’s characteristics as a source of

anomalies due to their changing abilities, preference and updating learning style. It handled learning time, score point, attention

interaction between tutor and leaner, understanding concept and feedback. Overall performance of the e- learning course by using DT

algorithm is improved by reducing error factor compared to the existing cloud computing. 70% of error is reduced comparing to the

existing.

Keywords— E-Learning, FAM-CART, DT, Kano Analysis, User Interface Design, Satisfaction.

I. INTRODUCTION

In recent years, e- learning (called electronic learning) is one of the common learning systems used by many people. A learning

system based on formalized teaching but with the help of electronic resources is known as E-learning. While teaching can be

based in or out of the classrooms, the use of computers and the Internet forms the major component of E-learning. E-learning can

also be termed as a network enabled the transfer of skills and knowledge, and the delivery of education is made to a large number

of recipients at the same or different times. Earlier, it was not accepted wholeheartedly as it was assumed that this system lacked

the human element required in learning. However, with the rapid progress in technology and the advancement in learning systems,

it is now embraced by the masses. The introduction of computers was the basis of this revolution and with the passage of time, as

we get hooked to smartphones, tablets, etc., these devices now have an important place in the classrooms for learning. Books are

gradually getting replaced by electronic educational materials like optical discs or pen drives. Knowledge can also be shared via

the Internet, which is accessible 24/7, anywhere, anytime.

1.1 Types of e – learning

There are a number of types of e-learning that depend on the amount of physical interaction. Entirely online e-learning occurs

without any face-to face interaction. Course work and materials are distributed electronically through email, websites, online

forums and/or CDs or DVD-ROMs. Combined learning uses a combination of Internet-directed instruction, as well as face-to-face

interaction [1]. Most traditional colleges and universities use combined learning as students learn in physical classrooms, with

instruction augmented by online lessons. For those learning for personal accomplishment, e-learning can also use a combination

of e-learning types, as they can be entirely self-directed, or they can use the assistance of an expert in their selected field.

1.2 Popularity of e-learning

The popularity of e-learning is evident from the way students, teachers and parents perceive this concept in this era. Every day, we

turn to the Internet for gaining information and Online courses have mushroomed like never before. Virtual classrooms are on the

rise and the number of students registering for such courses has shot through the roof. In addition, e-learning has found its place

in the heart of training incumbent personnel in the corporate sector.[2]

The International journal of analytical and experimental modal analysis

Volume XI, Issue XII, December/2019

ISSN NO: 0886-9367

Page No:2470

1.3 E- Learning advantages

E- learning does help to save cost

E –learning makes the education more available

E-Learning does make students more mobile.

E-Learning does make the whole learning process more entertaining.

1.4 Benefits and improvement in educational sector

1) Cost effective and time saving

E-learning courses do not demand students to be present in the classroom necessarily. Students can proceed with the

courses from the comfort of their homes or any place they deem convenient. This cuts down the money on travel and saves a lot of

time.

Let’s say an institute condenses 2900 hours of classroom training into:

600 hours of web-based training

500 hours of classroom training

300 hours of distance learning

This cuts the time spent on the training by about 52%. And the cost reduces considerably too. It is a win-win situation. [3]

Figure 1 Benefits of E- Learning

2) Tracking Course Progress becomes Easier

A well-implemented Learning Management System (LMS) makes tracking course progress more effective and easy. In

addition, LMS makes assessing students’ capabilities a piece of cake. Therefore, an e-learning system, which includes an LMS,

can prove to be quite effective in tracking learner’s progress.

3) E-learning provides ample room to be discreet

It so happens sometimes that a student lags behind in the class, while others find themselves quite in sync with whatever is

being taught. And that student feels shy about questioning the incomprehensible, since that would shower him with unwanted

attention. The whole situation ends up being extremely embarrassing. Such a situation can be completely avoided in e-learning

and any failure can be kept from getting out in the open. Therefore, the chances of being jeered at can be reduced.

This appeals a lot to every student, mostly co-workers. So, these are what attract students to e- learning.



ISSN NO: 0886-9367

Page No:2471

E-learning allows enough room for students to manage their tasks as and when they find convenient.

It also provides a step-by-step explanation procedure that caters to students of any level of comprehensive skills.

E-learning takes personal learning to a whole new level.

4) E-learning encourages sharing

E-learning provides students with a chance to share the knowledge acquired through online communities. A discussions

forum can add value to the learning procedure, by incorporating scope for fruitful collaboration and conversation. Sharing of

resources in e-learning is also an extremely healthy way for education to flourish.

5) The target audience base for e-learning is quite large

The two important aspects that act as barriers to learning are:

Location

Time-frame

1.5 Challenges in e –learning

Even though the concept of e-learning is set to create major waves in the education sector in the recent years, the challenges are

streaming in. Many organizations have embraced e-learning with open arms, but the problems amount to a staggering sky-high

heap when it comes to implementing e-learning at the school level.

Internet is still luxury

The Internet is still a luxury in many parts of the country. A vast majority of the Indian population resides in rural areas. The

lack of infrastructure in such areas gives rise to connectivity and accessibility issues. However, the Government of India has been

instrumental in removing such barriers by implementing various measures. Two schemes have been launched to aid in e-learning

implementation:

National Mission on Education through Information and Communication Technology (NMEICT)

National Program on Technology Enhanced Learning

These two schemes have been solely launched to implement ICT in video and web-based learning.

Figure 2 Challenges in E- Learning

E-learning does not cover a lot of certification courses

The certifications that come with conventional learning is somehow lost in the e-learning concept of education. The e-

learning courses do not cover a lot of certification courses that are recognized by colleges and universities across India or abroad.

This pulls the e-learning courses out of sync with any stream of school education. It would take some time to renovate the

conventional educational system. The traditional education methods have enlightened generations for decades now. Even though



ISSN NO: 0886-9367

Page No:2472

you might feel that they have overstayed their welcome, it has become increasingly difficult for us to overthrow tradition

completely and embrace newer methods of learning with open arms. However, renovation in the old-school methods has seen the

light of the day with technology entering the industry. But a complete makeover in education with the e-learning methods would

still require some time to establish itself. [4]

Not all learners are tech-savvy

Even though the e-learning courses are available in a wide range of platforms for learners to choose from, a basic

knowledge of how to operate those devices is imperative to benefit from the courses. And being a tech-savvy teacher becomes a

primary requisite. Therefore, before e-learning could be implemented, learners and educators need to be educated about the ins

and outs of technology to facilitate a smooth learning curve[5].

Lack of awareness

If a large part of the population isn’t aware of the amazing benefits that e-learning has to offer, then how can it be

expected to change the face of education in the coming years? Awareness plays a key role in making the proliferation of e-

learning a joy ride. With that lacking, the future becomes questionable. While the challenges pose an impending storm rocking the

e-learning ship violently, the numerous benefits calm the waves to a soothing cradle[6]. E-learning streams in like a shining ray of

hope, making education accessible for:

Anyone

Anywhere

Anytime

2. USER INTERFACE DESIGN

User Interface Design is the discipline of designing software interfaces for devices, ideally with a focus on maximizing efficiency,

responsiveness and aesthetics to foster a good user experience. UI design is typically employed for products or services that

require interaction for the user to get what they need from the experience. The interface should allow a user to perform any

required tasks to complete the function of the product or service.

An interface is a point of interaction between the user and the hardware and/or software they are using. UI design is the

skill employed to visualise the interface used to complete the task it is designed for. Good UI design facilitates making the

completion of tasks as frictionless as possible and increasing usability[7]. A UI Designer is someone who creates the user

interface based on a functional requirement and planned user experience using design standards and aesthetics to craft a certain

experience. UI Design can be done with pens and pencils, computer visualisation software, or built directly in code or materials.

The end results is an interface, or a simulation of one, that can be used to test, iterate and release a product or service.

2.1 Types of user interface design

There are different ways of interacting with computer systems which have evolved over the years. There are five main types of

user interface:

Command line (cli)

Graphical user interface (GUI)

Menu driven (mdi)

Form based (FBI)

Natural language (NLI)



ISSN NO: 0886-9367

Page No:2473

2.2 Advantages of Good User Interface

UI is also very important for an effective and positive learning experience. Generally, both of these go together and are inter-

dependent. A good website is one that’s easy to use. Here’s why a good UI is important for effective learning, and its advantages

for an LMS:

(i) It Attracts More Learners

The greater the number of learners who sign up for an eLearning course on an LMS, the greater its popularity. If a course has

only a few applications or sign-ups, then it’s an indication of something being amiss. Students who sign up with an LMS have

some preset goals and expectations from the course; if these are not met, then the course isn’t helpful.

(ii) Allows Better Interaction With The System

For a user to interact with the system effectively, the layout needs to be clean, concise, and easy to navigate. If the call-to-

action buttons and menus are placed clearly and the content is visible, then the user does not need to take too much effort to find

their way around the page. On the contrary, if the color scheme is too loud, or the font isn’t readable, or even if the site is slow to

load and doesn’t adjust itself quickly, a user can very soon get tired of making attempts at understanding it.

(iii)Fulfills the Educational Aim

The main aim of an eLearning course is that it should fulfill the educational aim of a user. Every student expects some

valuable returns from the course; a good UI makes it easier for users to focus on learning without having to bother about the look

or navigation.

(iv) Helps Course Creators and Admins Too

Whether the site admin or course creator/author manages the course, if the UI is well-made, then the process is easier. A

complicated UI and UX take up more time and effort to use, adding to the user's frustration and lessening productivity as well. A

great UI enables faster and easier creation, modification, addition, and deletion of courses, lessons, topics, quizzes, exams, and

feedback.

3. APPLYING DATA MINING TECHNIQUES FOR E- LEARNING PROBLEMS

E-learning (also referred to as web-based education and e-teaching), a new context for education where large amounts of

information describing the continuum of the teaching-learning interactions are endlessly generated and ubiquitously available.

This could be seen as a blessing: plenty of information readily available just a click away. But it could equally be seen as an

exponentially growing nightmare, in which unstructured information chokes the educational system without providing any

articulate knowledge to its actors. Data Mining was born to tackle problems like this. As a field of research, it is almost

contemporary to e-learning. It is, though, rather difficult to define. Not because of its intrinsic complexity, but because it has most

of its roots in the ever-shifting world of business. At its most detailed, it can be understood not just as a collection of data analysis

methods, but as a data analysis process that encompasses anything from data understanding, pre-processing and modelling to

process evaluation and implementation. It is nevertheless usual to pay preferential attention to the Data Mining methods

themselves. These commonly bridge the fields of traditional statistics, pattern recognition and machine learning to provide

analytical solutions to problems in areas as diverse as biomedicine, engineering, and business, to name just a few. An aspect that

perhaps makes Data Mining unique is that it pays special attention to the compatibility of the modelling techniques with new

Information Technologies (IT) and database technologies, usually focusing on large, heterogeneous and complex databases. E-

learning databases often fit this description[8].

Therefore, Data Mining can be used to extract knowledge from e-learning systems through the analysis of the

information available in the form of data generated by their users. In this case, the main objective becomes finding the patterns of

system usage by teachers and students and, perhaps most importantly, discovering the students’ learning behavior patterns.

Data mining and E-learning Aims to provide an up-to date snapshot of the current State of research and applications of

Data Mining methods in e-learning. The Cross-fertilization of both areas is still in its infancy, and even academic References are



ISSN NO: 0886-9367

Page No:2474

scarce on the ground, although some leading education-related Publications are already beginning to pay attention to this new

field. In order to Offer a reasonable organization of the available bibliographic information According to different criteria, firstly,

and from the Data Mining practitioner Point of view, references are organized according to the type of modeling Techniques used,

which include: Neural Networks, Genetic Algorithms, Clustering and Visualization Methods, Fuzzy Logic, Intelligent agents, and

Inductive Reasoning, amongst others. From the same point of view, the Information is organized according to the type of Data

Mining problem dealt with: clustering, classification, prediction, etc.

Finally, from the standpoint of the e-learning practitioner, we provide taxonomy of e-learning problems to Which Data

Mining techniques have been applied, including, for instance: Students’ classification based on their learning performance;

detection of Irregular learning behaviors; e-learning system navigation and interaction Optimization; clustering according to

similar e-learning system usage; and systems’ adaptability to students’ requirements and capacities.

4. DATA MINING FUNCTIONALITIES FOR E-LEARNING DOMAIN

In the e-learning domain, we are interested in managing mainly two groups of users: the learners as well as the learning

providers, whether private training companies, governmental organizations and local authorities providing training for their

employees or universities who aim to publish their courses and make them accessible online via the Internet. As for learners,

databases should store all personal details including name, age, gender, address, postcode, and educational-relevant details such as

qualifications. Moreover having information like work experience, career objectives, income range, previous courses taken and

courses of interest would be of great value to be able to predict future behavior of different classes of employed professional

people. Also other information such as personal interests and hobbies would be very valuable for data mining tool in order to

discover hidden patterns by building intelligent models based on the huge amount of data.

In our proposed work, we use data mining technique for detecting anomalies and proposed decision tree with fam-cart

for detecting content material , detecting e- learning course anomalies and estimating student satisfaction in web based learning

using kano analysis and principle of delone’s & McLean success model [9].

5. DETECTION OF COURSE ANOMALIES

E – Learning is recent learning system and used mainly in educational sector. E-learning adaptation has become the most

important method that facilitates access to the appropriate content. Adaptive approaches consist of reducing the problems of

incompatibilities between learner’s cognitive abilities and educational content’s difficulties. In some cases, the adapted curriculum

cannot meet learner's skills completely seen its incoherent structure, its unsuitable methodologies and sometimes its complexity.

Therefore, we need to measure the convenience of the content material to improve it and ensure learners’ satisfaction. In other

words, it is necessary to estimate its appropriateness to each learner.

The proposed work proceeded by decision tree (DT) algorithm which is a supervised data mining method. It helps to

predict the convenience of the proposed content material for learners. It consists of classifying learning material into two classes:

―good‖ if it is convenient, and ―anomaly‖ if not. To achieve that, we have used an intelligent agent called FAM – CART model.

Farm – cart is useful to tackle data classification problems. Online FAM network is useful for conducting incremental learning

with data samples, whereas the CART model prevails in depicting the knowledge learned explicitly in a tree structure. Advantage

of hybrid FAM – cart model is capable of learning incrementally while explaining its prediction with elicited from data samples. It

tracks learner’s behavior by collecting a set of attributes like score points, learning time, number of attempts, feedback of learner,

interactions with the tutor , learning rate, attention, new teaching methods and understanding concept.

Then calculates the predictive attribute by using the (DT) algorithm. The finding algorithm shows that the score is the

most crucial indicator gives us more information about the conformity of curriculum to learners, followed by learning time,

feedback and number of attempts, score points. We calculated a student satisfaction using e- learning adaptation. In our proposed

work estimating the student satisfaction using Delone &Mclean are applied for Kano two dimensional model. Defining attribute

for student satisfaction and dissatisfaction. Finally in our work, we enhance three technique for e – learning to make more efficient

comparing to before [10].



ISSN NO: 0886-9367

Page No:2475

Figure 3 Block diagram of proposed work

5.1 Data Mining For Detecting E -Learning Course Anomalies

In E-learning field, the objective of the majority of research works was to improve the learning process, by ensuring a

personalized learning path, enhancing learner’s performance. Moreover, also, proposing methods to adapt the curriculum to

learner’s abilities by predicting learners’ behaviors and preferences. They have focused on learner’s characteristics as a source of

anomalies due to their changing abilities, preferences and their updating learning style.

The main goal of these researches has consisted of improving learning and predicting the most attractive ways to ensure

relevant learning. Except, learning problems does not imply necessarily the incompatibility problems between learners’ abilities

and curriculum difficulties. In some cases, adaptation’s problems can due to the proposed content material. It can affect the

learning process and then blocks learners. [11] For example, incoherencies in course’s section or bad structure influence learners’

understanding. In this context, the purpose of our research is to conceive an intelligent system by using (DT) algorithm as a

supervised method that:

Detect learning content materials which constitute anomalies for learners by predicting their convenience to learners’

abilities.

Ensure a good quality of courseware, by inserting another one more convenient and adapted.

Classify the leaner’s capabilities like score points, learning time, attention, learning rate, understanding concept,

feedback of learners.

5.2 E-learning model for detecting curriculum anomalies

1) (DT) s algorithm:

The (DT) algorithm represents an effective method of decision-making. It represents a set of choices in the graphical tree

form. It consists of partitioning data into homogeneous groups as possible according to the predictive variable (decision variable).

It takes as input classified data, and we obtain as output a tree that looks like an orientation diagram. Each last node (leaf)

represents a decision class, and each node (internal) is a new test. Each leaf represents a decision belonging to a class. Data verify

all path test leading from the root to that leaf.

The (DT) algorithm has the advantage of being readable and fast to execute. It is an automatically computable representation

of supervised learning algorithms. We have two types:

The classification trees: consists of explaining or predicting the appurtenance of a set of objects (observations, individuals)

to a class, or category that is considered as a qualitative variable, based on quantitative or qualitative explanatory

variables.

The regression trees: are used to explain or predict the values taken by a quantitative dependent variable as a function of

quantitative or qualitative explanatory variables. The classification model is based on a set of recorded data called

dataset. It constitutes the input of learning algorithm. The generated model must fit the input data (training set).



ISSN NO: 0886-9367

Page No:2476

Moreover, predict the class label of newly recorded attributes (Test Set) that have never seen before. The objective is to build a

learning model (tree, graph) with a good generalization capability explained in the above figure 4

5.3 E-learning system for classifying content materials

The decision tree technique is widely used in e-learning systems seen its efficiency in predicting and classifying

learners’ performance. It can be considered as a sort of tracking learners to analyze their behaviors and the main objective remains

to provide a relevant learning curriculum. Also, it can be used as a mechanism to detect courses’ anomalies during learning. In this

context, our research aims essentially to detect problems can affect learners during the learning phase. Namely, problems in

course's structure, its methodologies or a high abstraction level compared to their cognitive skills. The objective is to find the

algorithm that allows the best and the precise classification of e-learning content (good, anomaly) by exploring some learners'

attributes.

To get information about the appropriateness of the content material, the proposed e-learning system consists of tracking

learner’s behavior in different learning steps. We have used an intelligent agent called FAM- CART. It collects attributes like

learning time, score unit, number of attempts, interactions with the tutor and learners’ feedback. This attributes collection

indicates if the educational content fits their understanding or anomalies are preventing the effective learning. The (FAM- CART)

classify the proposed learning object (LO) based on (DT) algorithm in the figure 4

Once registered in e-learning system, learners pass an assessment test to determine their appropriate level. We define

three levels, beginner, intermediate and advanced. The educational curriculum is composed of a short units’ collection helps them

to quickly understand and evaluate information through a test at the end of each unit. Our objective is to define if the content

material responds to their needs or not. Namely, measure its appropriateness in term of each unit. The (FAM-CART) collects

attributes which constitute the input (data set) of (DT) algorithm. [12]

Figure 4 Architecture of Our Classification System



ISSN NO: 0886-9367

Page No:2477

Learning time (LT): This attribute gives us more information about the time spent in learning and understanding (LO) in

term of each unit. When learners face problems in learning, they try to assimilate some concepts by communicating with

the tutor and looking for more courses to understand. So, the more learning period is long more it includes understanding

problems P

Score points (SP): represents the obtained score at the end of the unit test. It reflects the comprehension level of the

learned concepts.

Attention (A): represents the leaner’s capacity or concentration while listening the topic or lesson.

Some attempts (NA) represents the number of repetition times to validate the test unit. More the number of attempts is

high more it means that the course does not fit precisely the learners’ cognitive abilities.

Interaction with the tutor (I): represents the frequency of interaction with the tutor through chat and messages. When

learner cannot assimilate some notions and concepts, he tries to contact the tutor. The frequency of interaction between

them can reflect the existence of anomalies.

Feedback of leaner (F): represents learners’ feedback and their opinions about the structure and methodologies of (LO).

If learners’ are dissatisfied with the (LO), their feedback will be negative.

5.4 Building a attributes in DT algorithm

The obtained (DT) algorithm allows to detect the convenience of the (LO) to learners dynamically and classify it (good,

anomaly). Namely, it determines the importance of each attribute in the process of identifying course’s quality. It shows that the

essential attribute impacts strongly on the predictive decision is the score. This variable is continuous, so the algorithm defines

discretization threshold which permits the best partitioning. In our case, the chosen score threshold is 35. If learner’ score is

higher than 35, the (LO) is classified as good content and as pure leaves as shown in the (DT) Fig.3. In the opposite case, another

attribute (LT) is selected. It gives more information about the time spent in understanding the (LO). In this step, if the learner

exceeds the (LT) threshold, this attribute takes (No), it means that there may be some problems in the explanation methodology,

and then it will be directly classified as an anomaly. If not, the algorithm selects another attribute which is learner’s feedback to

get more information about its convenience to proceed in the proposed (LO). Indeed, if learner gives negative feedback, the

content unit will be selected as an anomaly. In the case of positive feedback, the algorithm chooses the (NA). This attribute allows

a decision to be made as to whether the (LO) meets learner’s needs or not. In particular, if he exceeds the (NA) allowed

(threshold).The content will be classified as an anomaly. Otherwise, the content is good.

Figure 5 decision tree for detecting anomalies



ISSN NO: 0886-9367

Page No:2478

The (DT) algorithm selects attributes according to its importance and its impact on learners' achievement. The obtained

algorithm shows that the score is the most influencing one giving the best splitting. Therefore, the course is classified as good

content if (S) exceeds >35. In general, the score reflects the success on the (LO), and having a good mark means understanding its

concepts and proves its relevance and then motivates learners to continue.

In the case of a score less than this threshold, and because the algorithm evaluates the content, not the learner, it takes

into account learning time as a second indicator of its appropriateness. Thus, the more learner respects the time devoted to

learning; more it reflects the relevance of the content and its methodologies. Learners cannot proceed if they do not like the

content, even it is interesting (F=negative). That is why the (DT) algorithm considers learners’ feedback as a third important factor

influencing the learning process.

The number of attempts to validate the (LO) is the last attribute affecting the learning process. The more learner does not

exceed the threshold allowed to succeed; more it reflects the existence of anomalies in the structure and the methodology of the

proposed (LO). We notice that interactions with the tutor attribute do not appear in the (DT), it is relied strongly on to the

learning time attribute, because when learner communicates with the tutor, he spends the allocated time for learning and

understanding unit, and then he can overtake the learning time threshold.

The (DT) algorithm consists of dividing recursively and most efficiently the learning set examples. This operation is

based on tests defined using the attributes until we obtain subsets containing (almost) as examples belonging to the same class.

The splitting criteria are determined so that the result of the partition at each branch is pure as possible. For this reason, the

entropy has been introduced as a good indicator to measure the impurity as shown in the following equation (1):

(1)

is the probability appearance of class 1 in the data set(d)

We also have another function that permits to measure the degree of heterogeneity of classes in all samples, and in any

position of the tree in construction. It defines the gain for a set D of examples and an attribute Q. it represents the difference

between the entropy before and after splitting (2):

(2)

5.5 FAM- CART Algorithm For Reterving Attributes

Once connected, learner chooses the appropriate (LO). When he starts learning in the selected one, the FAM-CART

initializes two essential attributes: Learning Time (LT=0) and the number of interactions with the tutor (I=0). The FAM-CART

launches counter calculating learning time (LT). At the same time, it counts the number of interactions with the tutor (I).



ISSN NO: 0886-9367

Page No:2479

Figure 6 flowchart for FAM-CART algorithm

For example, if learner communicates with him asking for help or explanation, the attribute (I) will be incremented. By finishing

the (LO), the FAM-CART saves the recorded values (LT) and (I) in the track.xml file. Then, students must pass test assessment to

evaluate the comprehension level and the efficiency of the proposed (LO).So, we have two cases:

• If he succeeds (S>=35 ): the agent saves the obtained (S), learner’s feedback (F), Interactions (I) and Number of attempt

(NA =0).

• Else, the agent increments the (NA) and learner must repeat the (LO). The agent does not save any information.

Calculating the error and improve performance of e- learning and finding the error values of system:

Root Mean Squared Error (RMSE)

RMSE =

Mean Absolute Error (MAE)

(MAE) =

Relative Absolute Error (RAE)

(RAE) =

Root Relative Squared Error (RRSE)

(RRSE)=



ISSN NO: 0886-9367

Page No:2480

6 ESTIMATING STUDENT SATISFACTION BY KANO ANALYSIS

Within conducted study Kano model questionnaire is used to understand students’ satisfaction with the web based

learning system. In order to define quality attributes for Kano model, five quality components of Delone and Mclean (D&M)

model have been used below table. The questions for each quality attribute of web based learning system, systems quality, service

quality, information quality, use, and net benefits, have been created. The responders have been asked about their mindset due to

functional and dysfunctional dimension of web based e learning system quality attribute.

For example, the offered answers in both cases, in accordance with Kano model, are as follows: I like it; it must be that

way; I am neutral; I can tolerate it; or I dislike it. The respondents have to choose one of the offered options (answers) for both

functional and dysfunctional dimension of the question. Due to the chosen pairs the reviewers may get an overview of the

students’ satisfaction of the web based learning system quality attributes

Using this model, quality attributes are classified into six categories are

1. Attractive quality attribute (A): an attribute that gives satisfaction if present but that produces no dissatisfaction if absent;

2. One-dimensional quality attribute (O): an attribute that is positively and linearly related to customer satisfaction—that is, the

greater the degree of fulfillment of the attribute, the greater the degree of customer satisfaction;

3. must-be quality attribute (M): the presence of these product/service attributes will not increase customers’ satisfaction level

significantly, while their absence will cause extreme dissatisfaction;

4. Indifferent quality attribute (I): an attribute whose presence or absence does not cause any satisfaction or dissatisfaction to

customers;

5. Reverse quality attribute (R): an attribute whose presence causes customer dissatisfaction, and whose absence results in

customer satisfaction;

6. Questionable quality attribute (Q): it means that it is not clear weather customers expect these attributes, since they gave

unusable responses due to misunderstanding the questions on the survey or making an error when filling out the

questionnaire.

Quality attributes Kano model questionnaires

System quality (i) Technical stability \ reliability of the system

(ii) User friendly interface

Service quality Available access to the system at any time quality \ quantity of e – instructional material

Use (i) Presence of audio\video recordings

(ii) Student of self-assessment possibilities

(iii) Mandatory test and e – assignment

(iv) Collaborative activities

(v) Presence \ existing of e- tutors

Net benefits Enhance learning with combination of web based and traditional learning model.

Table 1 kano model attributes are defined by Delone and MacLean



ISSN NO: 0886-9367

Page No:2481

Figure 7 Kano model 2d graph for functionality and dysfunctionality

Evaluation according to Customers’ Satisfaction Indexes Since the results of the analysis in the previous case are fuzzy, we do

here an effort to ―sharp‖ them slightly, throughout the further analysis Namely, instead of concerning must-be (K), one-

dimensional (L), and attractive (F) features, the responses of the customers are reduced here to two numbers: a positive number

that is the relative value of meeting this customer requirement (versus the competition)and a negative number that is the relative

cost of not meeting the customer requirement. These numbers are labeled as ―better‖ (1) and ―worse‖ (2) indexes and calculated in

the following way, that is, by

(3)

(4)

Better (or, satisfaction index) indicates how much customer satisfaction is increased by providing certain feature of a system

which is intended to be developed, while worse (or, dissatisfaction) indicates how much customer satisfaction is decreased by not

providing the feature. More precisely, the positive better numbers are indicative of the situation where, on average, customer

satisfaction will be increased by providing attractive and one-dimensional elements. The negative worse numbers are indicative of

the situation where customer satisfaction will be decreased if these one-dimensional and must-be elements are not included into

―exacted‖ blended/e learning system which designers, teachers, e-tutors, and so forth are intended to develop by meeting the

learners’ (customers’) expectations. Now, let us consider in the light of these two coefficients the results of the survey being

conducted here and try to create more specified picture of the customers’ expectations.

Kano model questionnaires (estimating student satisfaction)

Q1:technical stability \ reliability of the system

Q2: user friendly interface

Q3: quality \ quality of instructional materials

Q4:presence of audio \ video recording

Q5: collaborative activities

Q6: self – evaluation possibilities

Q7: mandatory exercise, test, essays.

Q8:combination of web based and traditional learning

Q9: presence\ existing of e- tutors

Q10: available access time of a system

Table 2 kano questionnaire

(i) Presence of audio/video recordings seems very important for the customers; that is, it implies must be requirement. Its

absence will cause consequently great dissatisfaction (the better index is the largest for Q4).



ISSN NO: 0886-9367

Page No:2482

(ii) Collaborative activities, quality/quantity of instructional materials, and user-friendly environment (Q5, Q3, and Q2) have

large better indexes, which means that their absence will also cause dissatisfaction among the users.

(iii) To the availability of the access to the system at any time, as well as technical stability/reliability of the systems (Q10

and Q1), the customers did not give high scores. This can be explained as something that they take for granted a priori.

Or, in other words, it is quite normal for them that these two conditions are present, so they do not think they require

special concerning. However, this statement should be taken with a certain dose of reserve.

(iv) Presence of e-tutor(s) is considered unimportant for the students (the smallest value of better index for Q9).

This could be explained by the fact that students are sufficiently familiar with information systems and that they do not

need e-tutor. Now, by taking into the consideration the negative indexes, the following can be observed.

(i) Absence of audio/video instructional materials causes dissatisfaction among the customers (the worse index absolutely

value is the largest for Q4). This is completely in accordance with the previous statements due to this feature.

(ii) Also, absence of e-learning system stability/reliability will imply customers’ great dissatisfaction. This is logical, even it

is not completely in accordance with the previous customers’ judgments about this feature.

(iii) The requirement that causes the lowest degree of dissatisfaction among users is not providing user friendly environment

(the worse index absolutely value is the smallest for Q2). It can be concluded that its presence is convenient, but its

absence will not cause excessive dissatisfaction.

(iv) The levels of dissatisfaction which can be caused by the absence of the remaining features are rather of equal level, which

implies that their absence will not extremely affect the customers’ needs.

Kano model categories appearance have been measured and some approximations have been done in order to make the

responses more meaningful. Also some additional analysis based on determination of ―better‖ and ―worse‖ indexes have been

made with the aim of reducing the fuzziness in observations as much as possible. Some two-dimensional graphical analyses have

been realized as well. These analyses result in ―shifting‖ some points to other more appropriate Kano categories or 2D graphic

quadrants, due to the researcher’s empirical point of view. It is to be noted that there is a scattering among the obtained results and

that this is to be reduced throughout repeating the questionnaire among another considerably larger target group(s) of students,

modifying the questions, and/or including some additional questions into the model.

However, to the designers of e-learning systems in blended environment should be recommended to combine different

analytical and/or stochastic methods in assessing degree of customers’ expectations and their level of satisfaction. Holistic

approach based on users’ satisfaction level and the appropriate measurement analysis should give support to the designers in

improving existing and designing new more attractive web based learning models in the contemporary educational blended

schemes.

And finally, speaking more generally, as a powerful communications and commerce medium, the Internet is a

communication and IS phenomenon that lends itself to a measurement framework (e.g., Kano and D&M models).Within the e-

commerce context, the primary system users are customers or suppliers rather than internal users. Customers (students/learners)

and suppliers (teachers/instructors) use the e-system for learning as well for buying or selling learning courses.

7 Efficiency analysis of e- learning system

7.1 Student’s satisfaction of the overall enhancement in e –learning course

In our proposed work, reducing the problem of incompatibilities between tutor and learner, learner cognitive abilities and

educational content difficulties. It also removes incoherent structures, unsuitable techniques or methodologies and it makes

sometimes more complexity. We proposed e – learning model for detecting anomalies and content material using decision tree

with FAM – Cart algorithm. From the proposed work, student satisfaction is improved comparing to the existing cloud computing.



ISSN NO: 0886-9367

Page No:2483

Table 3 student’s satisfaction in overall enhancement in e- learning course

Questionnaire

Student’s Satisfaction In Overall Enhancement In

E –Learning Course

Satisfied Average Dissatisfied

Q1 72.3 22.3 6.9

Q2 74.2 18.5 25.3

Q3 77.6 14.2 2.3

Q4 77 16.2 5.3

Q5 70.2 15.2 12

Q6 82.3 7.6 2.7

Q7 85.2 8.9 6.5

Q8 74.2 32.1 4.2

Q9 89.6 14.2 7.4

Figure 8 student satisfaction in overall enhancement in e –learning course

From the above figure represents student’s satisfaction in an e – learning course. We adopted data mining to rectify

problems in e- learning. We calculated customer (learner or student) learning time, score points, number of attempts, feedbacks

and interaction with tutor. We find content material is good or anomaly and new teaching methods which helps a student more

convenient to learn lesson patterns. These are improved student satisfaction comparing to the existing e –learning techniques.

Delone &Mclean are applied for Kano two dimensional model. Defining attribute, questionnaire model is to analysis

student satisfaction and dissatisfaction. Finally in our work, we proposed three technique for e – learning to make more efficient

comparing to before. Moreover, the percentage of student satisfaction is increased and unsatisfied, average terms are decreased.

Learning time, score points, interaction with the tutor of each student is improved. Then overall performance of each student is

improved.

7.2 Response Time

ELAPSED

TIME (min:sec)

RESPONSE TIME(ms)

DT WITH FAM –

CART

CLOUD

COMPUTING

ELARS

0.00 3000 3700 4100



ISSN NO: 0886-9367

Page No:2484

1.00 2900 3400 3900

1.25 2800 3250 3500

2.00 2300 2500 3000

3.00 3800 4100 4500

4.00 3200 3500 4000

5.00 2000 2200 3500

6.00 1500 2500 3000

7.00 1250 2000 2500

Table 4 Response Time of E- Learning System

From this figure shows response time of e- learning course adaptation. Response time refers to the amount of time

Application Server takes to return the results of a request to the user. The response time is affected by factors such as number of

users, number and type of requests submitted, and average think time.

Figure 9 response time

In this section, response time refers to the mean, or average response time. Each type of request has its own minimal

response time. However, when evaluating system performance, base the analysis on the average response time of all requests.

The faster the response time, the more requests per minute are being processed. However, as the number of users on the

system increases, the response time starts to increase as well, even though the number of requests per minute declines.

The response time of ELARS is high and it does not accept the new request. It decline for some time and overall the

system delayed due to the number of users and response time is increases. Cloud computing is sufficient for enhanced learning

process in e – learning system. Numbers of users increases and response time is low. The system is compact.

In our proposed work is a supervised approach and it have faster the response time, the more requests per minute are

being processed. However, as the number of users on the system increases, the response time is decreased as well, even though the

number of requests per minute processed. Overall system has low response time and it tackle problems of large number of users

with less response time.



ISSN NO: 0886-9367

Page No:2485

7.3 Performance ratio:

Algorithm PERFORMANCE

MAE RMSE RAE RRSE

Decision tree with fam –cart 0.12 0.24 0.47 0.65

Cc in e -learning 0.11 0.35 0.38 0.41

NAVIE - BASED 0.12 0.27 0.32 0.57

Table 5performance ratio

Figure 10 performance ratio

From above figure describes performance ratio of three algorithm depends on root mean square error(RMSE) ,

relative root square error (RRSE), relative absolute error(RAE) , mean absolute error(MAE). In cloud computing and naive based

it reduces the mean absolute error, root mean square error increasing, so that error will not reduce. Overall the performance of this

technique lags with some error while adopted in e – learning course.

Decision tree and farm – cart is a semi supervised approach. This learning algorithm used to learn model and apply to the

test set. So that relative root square error, root mean square is low comparing to the existing techniques. So error of the entire

system is reduced. Then the performance of the proposed algorithm is improved comparing to the existing techniques.

8 CONCLUSION

In this work, enhancing e – learning that Addressing the problem of unsuitable methodology in content material,

incompatibilities between learner’s cognitive abilities and educational content difficulties, learning time, teaching method and

interaction between tutor and learner in e- learning by decision tree (DT)with FAM-CART algorithm. DT is a semi supervised

data mining used to tackle e- learning problem. A FAM-CART is an intelligent technique used to solve the data classification

problem. The proposed content material is classified into two classes: ―good‖ if it is convenient to learner or ―anomaly‖ if not.

The majority of the research works were to improve e-learning field, by ensuring a personalized learning path and enhancing

learner’s performance. The proposed method is used to adapt the curriculum to learner’s abilities by predicting learner’s behaviour

and preferences. This work will be focused on learner’s characteristics as a source of anomalies due to their changing abilities,

preference and updating learning style. It handled learning time, score point, attention interaction between tutor and leaner,

understanding concept and feedback. Overall performance of the e- learning course by using DT algorithm is improved by

reducing error factor compared to the existing cloud computing. 70% of error is reduced comparing to the existing.



ISSN NO: 0886-9367

Page No:2486

REFERENCES

[1] Nazeeh Ghatasheh, 2015, ―Knowledge Level Assessment in e-Learning Systems Using Machine Learning and User Activity Analysis‖ International

Journal of Advanced Computer Science and Applications, Vol. 6, No. 4.pp:107-113. [2] Natasa Hoic-Bozic, Martina Holenko Dlab, & Vedran Mornar 2016, “Recommender System and Web 2.0 Tools to Enhance a Blended Learning Model‖,

IEEE Transactions on Education (Volume: 59 , Issue: 1 )

[3] Khairan D. Rajab,2018, ―The Effectiveness and Potential of E-learning in War Zones: An Empirical Comparison of Face to-Face and Online Education

in Saudi Arabia‖ IEEE Access, ( Volume: 6),pp: 6783 - 6794

[4] Aamer Hanif, Faheem Qaisar Jama, & Muhammad Imran,2018, ―Extending the Technology Acceptance Model for Use of e-Learning Systems by Digital

Learners‖ IEEE Access ( Volume: 6 ) pp: 73395 – 73404 [5] Rakesh Garg, Ramesh Kumar, and Sandhya Garg, 2019, ―MADM-Based Parametric Selection and Ranking of E-Learning Websites Using Fuzzy

COPRAS‖ IEEE Transactions on Education (Volume: 62, Issue: 1) pp: 11 – 18.

[6] Christophe Kiennert, Nathan De Vos, Manon Knockaertand& Joaquin Garcia-Alfaro, 2019, ―The Influence of Conception Paradigms on Data Protection in

E-Learning Platforms: A Case Study‖ IEEE Access (Volume: 7) pp: 64110 – 64119.

1. Behnam Faghih, Dr. Mohammad Reza Azadehfar & Katebi s.d, 2014, ―User Interface Design for E-Learning Software‖, the International Journal of Soft

Computing and Software Engineering [JSCSE].

2. Irina ionita, 2016, ―Data mining techniques for e-learning‖, Journal of Applied Computer Science & Mathematics (Vol: 10, issue: 2) pp: 26-31.

3. Srinualalintippayawong,ThanakritJulavanich&KanokwanAtchariyachanvanich, 2019, ―Reverse SQL Question Generation Algorithm in the DB Learn

Adaptive E-Learning System‖ IEEE Access (Volume: 7 ) pp: 54993 – 55004.

4. ManjeevanSeera, Chee Peng Lim &Shing Chiang Tan, 2019, ―A hybrid FAM-CART model for online data classification‖ Computational Intelligence. pp:

1–20. 5. Suhirman, Jasni Mohamad Zain, Haruna Chiroma F& Tutut Herawan,2014, ―Data Mining for Education Decision Support: A Review‖ iJET ‒ Volume 9,

Issue 6,

6. Fatiha Elghibari, Rachid Elouahbi & Fatima El Khoukhi, 2018, ―Data Mining for Detecting E-learning Courses Anomalies: An Application of Decision

Tree Algorithm‖, international journal on advance science engineering information technology, Vol: 8.



ISSN NO: 0886-9367

Page No:2487

Determine the User satisfaction on E-learning using the...

Documents

Transcript of Determine the User satisfaction on E-learning using the...