Identifying e learner’s opinion using automated sentiment analysis in e-learning

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
__________________________________________________________________________________________
Volume: 03 Issue: 01 | Jan-2014, Available @ http://www.ijret.org 479
IDENTIFYING E-LEARNER’S OPINION USING AUTOMATED
SENTIMENT ANALYSIS IN E-LEARNING
P.Bharathisindhu1
, S. Selva Brunda2
1
Assistant professor, Department of computer science, Navarasam arts and Science College for women, Arachalur.
2
Professor & Head, Department of CSE, Sasure College Of Engineering, Vijayamangalam
Abstract
E-learning is becoming more powerful and popular medium for learning through online. It is very important to identify the opinions of
the users of E-learning. E-learning commonly refers to teaching efforts propagated through the use of computers to impart knowledge
in a non traditional classroom environment. It is difficult to find the users mind whether they are satisfied with the E-courses.
Sentiment analysis helps to find about the users who are visiting the E-learning portals. Sentiment analysis refers to the use of natural
language processing text analysis and computational linguistics to identify and extract subjective information in source materials. It
aims to determine the attitude of speaker or a writer with respect to some topic or overall contextual polarity of a document.
Automated sentiment analysis helps to identify the sentiment of the users from the E-portal pages in which users generally browse on
a topic or area they are interested. This information can be used by E-learning system to identify the user’s emotional factors and can
analyse the user’s activity each time. The user’s sentiments towards a course can serve feedback for E-learning portals in online
learning environment. In the proposed E-learning system the automated sentiment analysis helps to analyse the E-portal visited pages
by the user. This helps the developer and educator to know where the learner concentrating on specified areas with any manual work
or any feed backs from the user. Here for automated sentiment analysis the Bayesian statistics classification, Naive Bayes classifier is
used.
Keywords: Automated Sentiment analysis, Bayesian classification, Naive Bayesian classification Support Vector Machine.
----------------------------------------------------------------------***--------------------------------------------------------------------
1. INTRODUCTION
The E-learning systems have recorded a rapid growth in the
last decade. The E-learning systems are very useful to students
and teachers. In the said systems, the content of E-learning
courses can be stored, arranged and distributed to the E-
learners in meaningful ways, and assessments can be
conducted and results presented to the learners. Some popular
learning systems are Black board learning system CCNet and
ALEKS.
Since the users of the E-learning portal are spread over the
globe it is unlikely difficult to evaluate and rate the
performance of E-learning Systems as compared with that of
traditional classroom approach. The task of sentiment analysis
is to classify the polarities like whether it is positive, negative
or neutral to check the sentiment. Further that may vary such
as happy, excited, joy, anger, sad, cry etc instead of single
sentences.
In proposed paper the automated sentiment analysis itself
collect the opinion of the E-learners how they are
concentrating on their courses. The tasks to perform here is
information extraction like discarding subjective information,
recognizing opinion oriented questions, summarization. With
the specified data the developers can concentrate on the
specific areas where the users can concentrate more and spent
more time on the topics. The opinions and reaction of the
learners of the proposed ideas are relevant in adopting the new
ideas in E-learning portals.
2. RELATED WORKS
In educational institution and in corporate, E-learning plays a
vital role. Recognition critical success factors (CSFs) and
performance evaluation measurements of E-learning can assist
educational institutions and organizations which apply these
electronic solution to understand how they gain goals of
adopting of E-learning and also perceive where need adaption,
improvement and change[1]. The social networks are analyzed
in this paper to represent the groups of students who have
similar contacts and interact in similar social circles [2]. In
many natural language processing tasks, Subjectivity and
sentiment classification has been used as a first phase to
generate more viable data. Research that done from these
additional layering ranges from question answering [3], to
conversation summarization and text semantic analysis [4]. In
such system there is no need of classrooms and coaching
system. For the last decade in E-learning courses evaluated
and analysed and rated only through feedbacks like ,email and
online questionnaire or manually at times. The sentiment
analysis plays a role for getting the feedback on overall rating
of the courses. Through sentiment analysis can assess how
much the E-learning courses are effective, and we can arrive

__________________________________________________________________________________________
an opinion on E-portal whether it is positive, negative or
neutral.
3. PROPOSED APPROACH
In proposed paper it makes E-learning more comfortable for
both learners and educators. For developers it makes very easy
to know the opinions of the learners through automated
sentiment analysis. To identify the sentiment in E-learning
portals and through natural language processing automated
sentiment analysis helps. There are so many parameters
available for gathering the opinions and employ the different
techniques and statistical methodology to evaluate sentiment
across the web. Automated analysis can be almost as good as
human analysis. In the proposed research the dataset is
collected with the various instances Gender, Age, Visited
pages, properties, news, groups, peer feed, labs and
challenges. The dataset which helps to find the automated
sentiment analysis of the E-learning portals. From the data
analysis the classification is done to classify the relevant data
which is useful for the developers. They can know the area of
interest and interest shown on the topics. From that they can
modify according to the users. Thus helps to increase the
number of users to the specific E-learning portals.
3.1 Acquirement
The collection of 100 users review from the website
Functionspace.org. The nature of our dataset is generalized.
The dataset is collected based on the various keyword like
visited pages, properties, discussions, news, groups,
challenges, labs and messages. The sentiment analysis used for
our dataset is
Table -1: Keywords of sentiment analysis used in the dataset
In discussion area the user can discuss about the topics and
interested areas. We collected the sentiment of the discussion
and it will help to motivate the developers to modify the
courses according to the user’s interest. Depends on the visited
pages of news the users can know the latest news available in
the E-learning market. With the analysis of that area the
developers can know how the users can concentrate on news.
4. CLASSIFYING SENTIMENT
Automated sentiment analysis reviews on-topic posts as the
user comes in and determines the sentiment of the post at the
sentence level, and aggregates positive, negative, or neutral
designation at the post level based on specified sentiment
keywords or phrases in a particular page or sharing of a page.
If a particular document or post touches multiple topics,
sentiment can be determined for each separate topic. Based on
the visited areas and on which subject the user can concentrate
more, the sentiment can be classified. The classification of
sentiment analysis can be done with gender, age, visited pages,
properties and subject etc. The properties used in our dataset
which shows about the users who are all visiting what type of
topics, whether they concentrating on latest activity, advanced
techniques or beginners topics.
Table 2: keywords of the properties used in the dataset
4.1 Machine Learning
Machine learning is a fundamental way that enables the
computer to have the intelligence. Its application which has
been used mainly the method of induction and the synthesis,
rather than the deduction has already reached many fields of
Artificial Intelligence [5]. The aim is to examine the sentiment
classification of topic based categorization. For this
classification we use Naïve Bayes classification is used. To
implement such algorithms in our dataset, we have to follow
the standard frame wok. Let {f1....fm} to be predefined set of
m features. Let ni(d) be the number of times fi occurs in
document d [6]. Then the document vector d: = (n1 (d), n2 (d)
. . . nm (d)) is used.
4.2 Bayesian Classification (BSC)
BSC predicts the E-learning courses how the users use the
courses according to the sentiment. Bayesian statistics offer a
most forward for making inferences. The prior information
about the process and the information about the process
contained in the data are two sources of information. Bayesian
classifiers use Bayes theorem, which says

)(
)
dp
cp
c
d
p
d
c
p j
j
j














Where p(cj | d) is the probability of instance d being in class
cj, The p(d | cj) is the probability of generating instance
dgiven class cj. The p(cj) is the probability of occurrence of
class cj. The (d) is the probability of instance d occurs.
4.3 Naive Bayes Classification
Positive P
negative N
Neutral Ne
LA LATEST ACTIVITY
KI K-INDEX
AT ADVANCED TOPICS
BT BEGINNERS TOPICS
IT INTERMEDIATE TOPICS

__________________________________________________________________________________________
Naive Bayes classification helps to assume the presence or
absence of the particular feature related feature that is related
to other feature in the given class. For example consider a fruit
an orange if it is orange in colour, round, and about 4” in
diameter. A naive Bayes classifier considers each of these
features to contribute independently to the probability that this
fruit is an apple, regardless of the presence or absence of the
other features.
The Naive Bayes classifiers is often represented as this type of
graph
Fig -1: Arrow shows the probability of classes.
Naive Bayes text categorization tends to perform optimum for
certain classes. A naive Bayes classifier assumes that the
presence (or absence) of a particular feature of a class is
unrelated to the presence (or absence) of any other feature. For
example, a fruit may be considered to be an apple if it is red,
round, and about 4" in diameter. Even if these features depend
on each other or upon the existence of the other features.
Table 3: Analysis of users concentrate on Properties
A naive Bayes classifier considers all of these properties to
independently contribute to the probability that this fruit is an
apple. Depending on the precise nature of the probability
model. Naive Bayes classifiers can be trained very efficiently
in a supervised learning setting. In many practical
applications, parameter estimation for naive Bayes models
uses the method of maximum likelihood.
Fig -2: numbers of visited pages of properties.
Here the properties are taken as sample for classification in
regards with subjects. The graphical representation of the
diagram shows that which age group were concentrating on
which properties of the topics. For example here the age group
between 29-39 were more concentrated on Intermediate level
topics, this shows the users of this age group concentrated
more on the specific topics of intermediate level. So the
developers come to know from the sentiment analysis of these
areas helps to concentrate on the users concentrated areas.
Fig 3: Numbers of users classified on their age
The histogram describes about the users on their age specifies
the age group between 20-40 were concentrate more on the E-
portal.
An advantage of the naive Bayes classifier is that it requires a
small amount of training data to estimate the parameters
(means and variances of the variables) necessary for
classification. Because independent variables are assumed,
only the variances of the variables for each class need to be
determined and not the entire covariance matrix.
5. SUPPORT VECTOR MACHINE
To analyze data and recognize pattern used for classification
and regression analysis
W= i iii xy
PROPERTIES PHY MATH CS
AT 9.0 5.0 7.0
IT 4.0 11.0 10.0
KI 4.0 7.0 1.0
LA 14.0 4.0 9.0
BT 4.0 14.0 12.0
TOTAL 35.0 41.0 39.0
cj
P(d1|cj)
p(d2|cj)
cj)
P(dn|cj
)

__________________________________________________________________________________________
Support vector machine (SVM) helps to find the similar hyper
plane between the classes. It helps the developers to enhance
the E-learning courses accordingly that make a updating
regularly on the same.
6. RESULTS
The result of the proposed system will helps the developer and
educator to identify the most concentrated pages in E-learning
portals. It determines the sentiment of the E-courses in
phrases. Also we can judge the E-learning portals .All the
sentiments are classified with Bayes rule. With the classified
data we can access the most accessed pages by the user.
Finally automated sentiment analysis shows the E-courses
highlighted the most frequently accesses courses. This
motivates the users that in which area it highly focussed by the
other user and also helps to know the current trends. It mainly
helps the developers to develop the further ideas. And it helps
to modify the courses accordingly. Give the fruit full
information on most visited pages and remove the untapped
pages and help to enhance E-portal developments. And also it
helps the developers to maintain their E-portals regularly.
REFERENCES
[1]. Introduce critical success factors (CSFs) of elearning for
evaluating e-learning implementation success. Published in:
Educational and Information Technology (ICEIT), 2010
International Conference on (Volume:1 )Page(s):V1-224 -
V1-228 E-ISBN : 978-1-4244-8035-7 Print ISBN:978-1-4244-
8033-3
[2]. H. Yu and V. Hatzivassiloglou. Towards answering
opinion questions: Separating facts from opinions
andidentifying the polarity of opinion sentences. In EMNLP,
pages 129{136, Sapporo, Japan, 2003.
[3]. J. Wiebe and R. Mihalcea. Word sense and subjectivity. In
Proceedings of the Annual Meeting of the Association for
Computational Linguistics, Sydney, Australia, 2006.
[4]. A Study and Application on Machine Learning of
Artificial Intelligence, Published in:Artificial Intelligence,
2009. JCAI '09. International Joint Conference Page(s): 272 –
274
[5]. Thumbs up? Sentiment Classiﬁcation using Machine
Learning Techniques Bo Pang and Lillian Lee Department of
Computer Science ,Cornell University Ithaca, NY 14853 USA
, Shivakumar Vaithyanathan ,IBM Almaden Research
Center,650 Harry Rd. San Jose, CA 95120 USA.
[6]. Stanley Chen and Ronald Rosenfeld. 2000. A survey of
smoothing techniques for ME models. IEEE Trans. Speech
and Audio Processing, 8(1):37–50.
[7]. Web Data Mining - Exploring Hyperlinks, Contents and
Usage Data Chap 11: Opinion Mining, 2nd Edition, July
2011.
[8]. Sentiment Analysis and Opinion Mining (Introduction and
Survey), Morgan & Claypool, May 2012.
[9]. Online Web Sentiment Analysis on Campus Network
Published in Computational Intelligence and Design (ISCID),
2011 Fourth International Symposium on (Volume:
2).Exploiting Combined Multi-level Model for Document
Sentiment Analysis Published in: Pattern Recognition (ICPR),
2010 20th International Conference.
BIOGRAPHIES
Assistant professor, Department of
computer science, Navarsam arts and
science college, Erode
Professor & Head, Department of CSE,
Sasure College Of Engineering,
Vijayamangalam.

Identifying e learner’s opinion using automated sentiment analysis in e-learning

More Related Content

What's hot (19)

Viewers also liked (20)

Similar to Identifying e learner’s opinion using automated sentiment analysis in e-learning (20)

More from eSAT Publishing House (20)

Recently uploaded (20)

Identifying e learner’s opinion using automated sentiment analysis in e-learning