In today’s social networking era, if one has to make decision about any product, service or individual performance, the availability of various comments, suggestions, ratings, and feedbacks are abundant. The required decision support data can be collected through different sources of Medias like newspapers, blogs, and discussion forums and from internet too. So surely, it leads to the selection of best product, service or individual if it is analyzed efficiently. In leading and competitive world, this is huge and practical need of industries, organizations to empower their qualities. In the recent years, the significant study is done in the field of sentiment analysis. However, the earlier work focused the implementation and evaluation of individual sub technique of sentiment analysis. Though these implementations produces significant results of sentiment or opinion analysis, the trust of decision makers is still in dangling to accept the results of such analysis. In this paper, initially, we have been described the brief review about the sentiment or opinion analysis system. Then the details are provided about the design and about how to build an automated opinion discovery system to enhance performance of sentiment or opinion analysis based on feature extraction sentiment analysis sub technique, natural language processing and data mining techniques in an integrated way

1. Poster Paper
Proc. of Int. Conf. on Advances in Information Technology and Mobile Communication 2013
Design of Automated Sentiment or Opinion Discovery
System to Enhance Its Performance
M.A.Jawale1, Dr.D.N.Kyatanavar2, and A.B.Pawar3
1
S.R.E.S.College of Engineering, Kopargaon, Maharashtra & JJTU, Rajasthan, India
Email: jawale.madhu@gmail.com
2, 3
S.R.E.S.College of Engineering, Kopargaon, Maharashtra & JJTU, Rajasthan, India
Email: {kyatanavar, anil.pawar1983}@gmail.com
Abstract— In today’s social networking era, if one has to make
decision about any product, service or individual performance,
the availability of various comments, suggestions, ratings,
and feedbacks are abundant. The required decision support
data can be collected through different sources of Medias like
newspapers, blogs, and discussion forums and from internet
too. So surely, it leads to the selection of best product, service
or individual if it is analyzed efficiently. In leading and
competitive world, this is huge and practical need of industries,
organizations to empower their qualities. In the recent years,
the significant study is done in the field of sentiment analysis.
However, the earlier work focused the implementation and
evaluation of individual sub technique of sentiment analysis.
Though these implementations produces significant results
of sentiment or opinion analysis, the trust of decision makers
is still in dangling to accept the results of such analysis. In
this paper, initially, we have been described the brief review
about the sentiment or opinion analysis system. Then the
details are provided about the design and about how to build
an automated opinion discovery system to enhance
performance of sentiment or opinion analysis based on feature
extraction sentiment analysis sub technique, natural language
processing and data mining techniques in an integrated way.
products on commercial websites and express their views on
almost anything in discussion forums and blogs, and on
social network sites. Now, if one wants to buy a product, one
is no longer limited to asking one’s friends and families
because there are many user reviews on the websites. For a
company, it may no longer need to conduct surveys or focus
groups in order to gather consumer opinions about its
products and those of its competitors because there is a
plenty of such information publicly available. However,
finding opinion sites and monitoring them on the internet
can still be a formidable task because there are large numbers
of diverse sites, and each site may also have a huge volume
of opinionated text. In many cases, opinions are hidden in
long forum posts and blogs [3].
It is difficult for a human reader to find relevant sites,
extract related sentences with opinions, read them, summarize
them, and organize them into usable forms [20]. Thus,
automated opinion discovery and summarization systems are
needed.
Sentiment analysis is not a single task, but it is a multifaceted problem [3] containing many sub-problems such as
object identification, feature extraction and synonym
grouping, opinion orientation classification and integration.
Survey reported in [5] covers techniques and approaches
that promise to directly enable opinion-oriented informationseeking systems. Here, focus is on methods that seek to
address the new challenges raised by sentiment-aware
applications, as compared to those that are already present
in more traditional fact-based analysis and given requirement
about building an integrated system that tries to deal with all
the multi-faceted problems altogether.
Index Terms— Feature Extraction, Opinion Mining, Part-ofSpeech, Sentiment Analysis, Subjective Classification.
I. INTRODUCTION
Sentiment analysis or opinion mining is the computational
study of opinions, appraisals, and emotions toward entities,
events and their attributes. In the past few years, it has
attracted a great deal of attentions from both academia and
industry due to many challenging research problems and a
wide range of applications[2].Opinions are important because
whenever we need to make a decision we want to hear others’
opinions. This is not only true for individuals but also true
for organizations.
However, there was almost no computational study on
opinions before the invention of web technologies because
there was little opinionated text available. In the past, when
an individual needed to make a decision, he/she was required
to ask for opinions from friends and families. When an
organization wanted to find opinions of the general public
about its products and services, it used to conduct surveys
and focus groups. However, with the explosive growth of the
social media contents on the websites in the recent past, the
world has been transformed. People can now post reviews of
© 2013 ACEEE
DOI: 03.LSCS.2013.2.79
II. RELATED WORK
The research in the field of sentiment analysis was started
with sentiment and subjectivity classification, which treated
the problem as a text classification problem. Reference [5]
covered that traditionally; text categorization seeks to
classify documents by topic. There can be many possible
categories, the definitions of which might be user- and
application dependent; and for a given task, we might be
dealing with as few as two classes (binary classification) or
as many as thousands of classes (e.g., classifying documents
with respect to a complex taxonomy). In contrast, with
sentiment classification often have relatively few classes
(e.g., “positive” or “3 stars”) that generalize across many
48

2. Poster Paper
Proc. of Int. Conf. on Advances in Information Technology and Mobile Communication 2013
domains and users. In addition, while the different classes in
topic-based categorization can be completely unrelated, the
sentiment labels that are widely considered in previous work
typically represent opposing (if the task is binary
classification) or ordinal/numerical categories (if classification
is according to a multi-point scale).
Sentiment classification identifies whether an opinionated
document (e.g., product reviews) or sentence expresses a
positive or negative opinion. Subjectivity classification
determines whether a sentence is subjective or objective [12].
It states that many real-life applications, however, require
more detailed analysis because the user often wants to know
what the opinions have been expressed on [2]. For example,
from the review of a product, one wants to know what features
of the product have been praised and criticized by consumers.
Let us use the following review segment on iPhone as an
example to introduce the general problem (a number is
associated with each sentence for easy reference):
“(1) I bought an iPhone 2 days ago. (2) It was such a nice
phone. (3) The touch screen was really cool. (4) The voice
quality was clear too. (5) However, my mother was mad with
me as I did not tell her before I bought it. (6) She also thought
the phone was too expensive, and wanted me to return it to
the shop. … “
The question is: what we want to extract from this review?
The first thing that we may notice is that there are several
opinions in this review. Sentences (2), (3) and (4) express
three positive opinions, while sentences (5) and (6) express
negative opinions. Then we also notice that the opinions all
have some targets on which they are expressed. The opinion
in sentence (2) is on iPhone as a whole, and the opinions in
sentences (3) and (4) are on the “touch screen” and “voice
quality” features of iPhone respectively. The opinion in
sentence (6) is on the price of iPhone, but the opinion/emotion
in sentence (5) is on “me”, not iPhone. This is an important
point. In an application, the user may be interested in opinions
on certain targets, but not on all (e.g., unlikely on “me”).
Finally, we may also notice the sources or holders of opinions.
The source or holder of the opinions in sentences (2), (3) and
(4) is the author of the review (“I”), but in sentences (5) and
(6) it is “my mother”.
Reference [17] illustrates, with this example in mind, we
can define sentiment analysis or opinion mining. It starts
with the opinion target.
Object and feature: In general, opinions can be expressed
on any target entity, e.g., a product, a service, an individual,
an organization, or an event. We use the term object to denote
the target entity that has been commented on.
An object can have a set of components (or parts) and a
set of attributes (or properties), which we collectively call the
features of the object [14]. A particular brand of cellular phone
is an object. It has a set of components (e.g., battery and
screen), and also a set of attributes (e.g., voice quality and
size), which are all called features.
Minqing Hu and Bing Liu [20] concluded that an opinion
can be expressed on any feature of the object and also on the
object itself. For example, in “I like iPhone. It has a great
© 2013 ACEEE
DOI: 03.LSCS.2013.2.79
touch screen”, the first sentence expresses a positive opinion
on “iPhone” itself, and the second sentence expresses a
positive opinion on its “touch screen” feature.
Opinion holder: The holder of an opinion is the person
or organization that expresses the opinion. In the case of
product reviews and blogs, opinion holders are usually the
authors of the posts. Opinion holders are more important in
news articles because they often explicitly state the person
or organization that holds a particular opinion.
Opinion and orientation: An opinion on a feature f (or
object o) is a positive or negative view or appraisal on f (or o)
from an opinion holder. Positive and negative are called
opinion orientations.
In reference [1], it is discovered that with these concepts
in mind, we can define a model of an object, a model of an
opinionated text, and the mining objective, which are
collectively called the feature-based sentiment analysis
model.
The general sentiment analysis model terminologies and
their descriptions can be found in [2]. Generally, Opinion can
be of the two types namely, direct and indirect opinion.
Additionally, opinion can be of type explicit and implicit
opinion. An explicit opinion on feature is an opinion
explicitly expressed on f in a subjective sentence. An implicit
opinion on feature is an opinion on feature implied in an
objective sentence. The following sentence expresses an
explicit positive opinion: “The voice quality of this phone is
amazing.”
The following sentence expresses an implicit negative
opinion: “The earphone broke in two days.” Although this
sentence states an objective fact, it implicitly indicates a
negative opinion on the earphone. In general, objective
sentences that imply positive or negative opinions often
state the reasons for the opinions.
In practice, not all five pieces of information in the
quintuple described in [2] above need to be discovered for
every application because some of them may be known or
not needed. For example, in the context of online forums, the
time when a post is submitted and the opinion holder are all
known as the site typically displays such information.
In the same sense, reference [4] focused on one type of
opinion sources, customer reviews of products and proposed
a novel visual analysis system to compare consumer opinions
of multiple products. To support visual analysis, they
designed a supervised pattern discovery method to
automatically identify product features from Pros and Cons
in reviews of format. A friendly interface is also provided to
enable the analyst to interactively correct errors of the
automatic system, if needed, which is much more efficient
than manual tagging.
The tasks of feature-level opinion mining usually include
the extraction of product entities from product reviews, the
identification of opinion words that are associated with the
entities, and the determining of these opinions’ polarities
(e.g., positive, negative, or neutral). In recent years, several
approaches have been proposed such as rule-based and
statistical methods on this subject, but few attentions have
49

3. Poster Paper
Proc. of Int. Conf. on Advances in Information Technology and Mobile Communication 2013
opinions and there are even more expressions (possibly
unlimited) that can convey these concepts. However, little
in-depth study has been done on many of them.
With the proliferation of social networking and ecommerce the information contained in the opinions/reviews
expressed by the people has grown by leaps and bounds.
Reference [18] presents an opinion search engine system
that incorporates two novel opinion mining algorithms. The
opinions are based on features and the orientation of these
opinions is also largely based on the features rather than a
product as a whole. People seem to like/dislike a specific
product because of some feature associated with the product.
The proposed framework not only classifies a review as
positive or negative, but also extracts the most representative
features of each reviewed item, and assigns opinion scores
on them.
Feature extraction and synonym grouping, they remain
to be very challenging as studied by Bing Liu [3]. Object
extraction is probably the easiest because many existing
information extraction algorithms can be applied. Integration
and matching of all 5 pieces of information required for
sentiment analysis in the quintuple as given in [2] is still
lacking, which is probably not surprising as the research
community likes to focus on individual sub-problems. This
leads us to the question of sentiment analysis accuracy, i.e.,
what is the accuracy of the current state-of-the-art algorithms?
This question is not easy to answer because there are so
many sub-problems.
The proposed research work intends on real-life
applications, to provide a completely automated solution.
The key of the proposed research work is to fully understand
the whole range of issues and pitfalls in sentiment analysis,
cleverly manage them, and determine what portions can be
done automatically and what portions need human assistance.
Beyond what have been discussed so far, it is also important
to deal with the issue of opinion spam (e.g., fake reviews).
Opinion spam refers to writing fake or bogus reviews
that try to mislead users or automated systems by giving
untruthful positive and /or negative opinions in order to
promote some target objects and /or to damage the
reputations of some other objects [8]. Detecting such spam
is needed because it can make sentiment analysis useless for
decision making process. There is a real and huge need in the
industry for such services to be implemented.
This system aims to find what people like and dislike about
a given product. Therefore how to find out the product
features that people talk about is an important step[27].
However, due to the difficulty of natural language
understanding, some types of sentences are hard to deal
with as stated in [8, 25]. Let us see some easy and hard
sentences from the reviews of a digital camera: “The pictures
are very clear.”, “Overall a fantastic very compact camera.”
In the first sentence, the user is satisfied with the picture
quality of the camera, picture is the feature that the user talks
about. Similarly, the second sentence shows that camera is
the feature that the user expresses his/her opinion.
While the features of these two sentences are explicitly
been paid to applying more discriminative learning models
to achieve the goal.
On the other hand, little research work has evaluated their
algorithms’ performance for identifying intensifiers, entity
phrases and infrequent entities [16]. This work particularly
adopts the Conditional Random Fields (CRFs) model to
perform the opinion mining tasks. Relative to related
approaches, it has not only highlighted the algorithm’s ability
in mining intensifiers, phrases and infrequent entities, but
also integrated more elements in the model so as to optimize
its training and decoding process.
Hong Liu [10] proposed a model about internet public
opinion hotspot detection and analysis, the main technique
of categorization proposed was text categorization.
According to the text properties of internet public opinion,
introduced Vector Space Model to express text opinion. Text
corpora are chosen from some new websites. It perform Kmeans clustering and Support Vector Machine (SVM)
classifier on the documents, the experimental result shows
that the efficiency and effectiveness of such method. Though,
the use of Data mining techniques is introduced but
refinement for each step of the approach proposed above is
needed and further issues are not addressed; mainly, Dynamic
monitoring technology is in demand which can monitor the
web sites to detect change in time; Data cleaning is timeconsuming and labor-intensive; Web content analysis cannot
stop at word frequency analysis because sometimes the result
is poly-semantic.
Moreover, it addresses a need for novel techniques that
will summarize and analyze the relevant information in a
principled and systematic way.
Reference [19] anticipates the introduction of a
collaborative framework that will further advance the state of
the art and establish new targets for the next decade.
Contradiction Analysis can possibly be the most demanding
field for such a framework, as it utilizes most of the opinion
mining methods, and at the same time defines its problems
on data of various types, ranging from opposite sentiments
to conflicting facts.
This discussion gives us a good clue of the main tasks
involved and technical challenges in sentiment analysis.
The research community has studied individual sub
problems of the sentiment analysis [6 and 7]. The most well
studied sub problem is opinion orientation classification
(i.e., at the document level, sentence level and feature level).
Hui Wang, Jiansheng Chen et al. [11] presents a set of
language patterns, which is composed of 22 rules, to extract
two-noun phrases from customer reviews. Additionally,
language rules can extract useful product features that human
taggers fail to annotate as specified in [13]. The experimental
results of [11] indicated that the accuracy of the classifiers
benefits from the increasing of the text’s length and also
varies.
The existing reported solutions are still far from required.
The main issue is that the current studies are still coarse. Not
much has been done on finer details. For example, on opinion
classification, there are many conceptual rules that govern
© 2013 ACEEE
DOI: 03.LSCS.2013.2.79
50

4. Poster Paper
Proc. of Int. Conf. on Advances in Information Technology and Mobile Communication 2013
mentioned in the sentences, some features are implicit and
hard to find. For example, “While light, it will not easily fit in
pockets.” This customer is talking about the size of the camera, but the word “size” is not explicitly mentioned in the
sentence. To find such implicit features, semantic understanding is needed, which requires more sophisticated techniques. However, implicit features occur much less frequent
than explicit ones. Thus, mostly existing study on opinion
analysis focus on finding features that appear explicitly as
nouns or noun phrases in the reviews. To identify nouns/
noun phrases from the reviews uses the part-of-speech tagging.
Opinion mining suffers from several different challenges,
such as determining which segment of text is opinionated,
identifying the opinion holder, determining the positive or
negative strength of opinion [15]. Opinion mining is
concerned with the human reviews, emotions and sentimental
discussion. Everyone has their own perception and concern
about a particular problem, issue, or topic. Opinionated text
may be fake, irrelevant and or ambiguous information.
Opinions are far harder than facts to describe. Opinion
sources are typically informally written and highly diverse.
Here, it gives clear indication about findings of opinion
from the existing research work involved for opinion analysis.
Their main contribution is to identify either opinion is
positive, negative or neutral based on explicit opinion and its
features identification only. A very little focus is given on
implicit opinion and its implicit features too as described in
above reviews of the digital camera. So the proposed research
work will focus on identifying explicit as well as implicit
opinions through their explicit and implicit features
analysis.
opinion data is available in very large amount and from this
huge data, extracting useful and effective information is a
challenging task [23, 24]. As Data mining (DM) techniques
are useful to do such kind of tasks effectively. So in the
proposed research work, it intends to make use of DM
techniques for the opinion analysis.
Even various natural language processing techniques are
efficient for fast document & its content processing. Thus,
to make the proposed research model more accurate, efficient,
these techniques will be suitable.
It concludes the objective of the proposed research work
is: “To build an automated opinion discovery system to
enhance performance of sentiment or opinion analysis based
on feature extraction sentiment analysis sub technique, natural
language processing and data mining techniques in an
integrated way.”
IV. SYSTEM MODEL
The motivation in the research work is to develop an
automated opinion or sentiment analysis system that will not
only mine the opinions but will also extract useful information
related to the item’s features and use it to rate them as positive,
neutral, or negative [9]. This feature based opinion mining
will help the user to focus on the features of the opinion/
product he/she is interested in.
Fig.1 gives an architectural design for automated
sentiment or opinion analysis system The system performs
the opinion analysis in three main steps: Data collection
and its preprocessing, opinion mining through opinion
mining engine and generation of opinion status.
The inputs to the system are in the form of datasets where
dataset will include product name and an entry page for all
the reviews of the products. The output is the opinion status
of the reviews. i.e. positive, negative or neutral opinion about
the product.
Given the inputs, the system first downloads all the
reviews, and puts them in the dataset for further processing.
Following section briefly outline the main components of
the proposed system as shown in Fig.1
Data preprocessing
The data from the dataset is preprocessed so as to set the
data in the format which is acceptable to the data processing
techniques .The outcome of this step will give us formatted
dataset which will be the input for the opinion miming engine.
As studied in literature review, most of the existing dataset or
database files available for opinion mining are in the form of
web pages formats [8], either in HTML or XML tags. So before
opinion mining, it is necessary to filter out such tags to get
opinion dataset only. This step is also aimed towards such
data preprocessing before to set the data in the format which
is acceptable to the further data processing techniques
Opinion Mining Engine
This is the main component of the proposed system model.
It further consists of two major steps of its
computation namely, Feature Extraction and Opinion
Direction Identification. The illustration of these steps is
III. OBJECTIVES
As stated earlier Sentiment analysis is multi-faceted
problem thus, there is need to build integrated system that
tries to deal with all multi-faceted problem altogether. To
address this, the main objectives of this proposed research
work are enlisted below:
The first aim is to find the key features i.e. explicit and
implicit features about an object that are talked about in
multiple reviews and their analysis. So it means that the
proposed opinion analysis system will concentrate on the
feature extraction to achieve effective and useful opinion or
sentiment analysis.
It leads to the requirement of integrated system that will
deal with these multi-faceted issues such as object
identification, feature extraction and synonym grouping,
opinion orientation, classification and integration in the
integrated fashion so that it will do sentiment or opinion
analysis effectively using explicit as well as implicit features
of opinion.
At the same time, the most of existing opinion discovery
systems are developed using Artificial Neural Network, SVM,
Soft-Constraints and Entropy Model, etc. techniques [21,
22]. Today, if one wants to do effective opinion analysis, the
© 2013 ACEEE
DOI: 03.LSCS.2013.2.79
51

5. Poster Paper
Proc. of Int. Conf. on Advances in Information Technology and Mobile Communication 2013
sentence. The output of this step is status of the reviews. i.e.
positive, negative or neutral opinion about the product.
V. USEFULNESS
There is a real and huge need in the industries for accurate
reviews for their products to improve customer satisfaction.
In IT industries, the proposed system will be useful to reduce
the issue of opinion spam. The practical need and the
technical challenges will keep the field vibrant and lively.
Government intelligence is another application that has been
considered. For example, it has been suggested that one could
monitor sources for increases in hostile or negative
communications.
Sentiment-analysis technologies for extracting opinions
from unstructured human-authored documents would be
excellent tools. Besides reputation management and public
relations, one might perhaps hope that by tracking public
viewpoints, one could perform trend prediction in sales or
other relevant data. Sentiment or opinion analysis would be
the basis for the creation of automated websites.
VI. IMPLEMENTATION DETAILS
Figure.1 Architectural Design for Automated Sentiment or Opinion
Analysis System
given in next session:
The feature extraction, which is the first and foremost
task of the proposed work, it extracts “frequent” features
that a lot of people have expressed their opinions on in their
reviews, and then finds those infrequent ones. The opinion
direction identification takes the generated features and
states the opinions about the feature through the opinion
status component of the system.
In this step, to perform part of speech tagging, use of the
NLProcessor like linguistic parser is proposed, which parses
each sentence and yields the part-of-speech tag of each word
(whether the word is a noun, verb, adjective, etc) and identifies
simple noun and verb groups (syntactic chunking). It will
help to identify the expressed opinion is implicit or explicit
type. The next step is to find features that people are most
interested in. In order to do this, here proposed system will
use association rule mining to find all frequent itemsets. Not
all frequent features generated by association mining are
useful or are genuine features. There are also some
uninteresting and redundant ones. Feature pruning aims to
remove these incorrect features. Opinion words are words
that people use to express a positive or negative opinion.
Observing that people often express their opinions of a
product feature using opinion words that are located around
the feature in the sentence, we can extract opinion words
from the review dataset using all the remaining frequent
features (after pruning). After opinion features have been
identified, one can determine the semantic orientation (i.e.,
positive or negative) of each opinion sentence. This consists
of two steps: (1) for each opinion word in the opinion word
list, there is need to identify its semantic orientation and (2)
then decide the opinion orientation of each sentence based
on the dominant orientation of the opinion words in the
52
© 2013 ACEEE
DOI: 03.LSCS.2013.2.79
The proposed design architecture simulation work is
already started using Python and it is in its initial stage. Even
visualization of same collected dataset is done through
WEKA data mining tool. The data collection, its
preprocessing details are given below.
Review Data Collection
The required dataset is collected from the internet which
is freely available for the research study. This dataset is a
subset of the opinion mining datasets released by Dr. Bing
Liu’s group from University of Illinois at Chicago. Their
dataset is available from http://www.cs.uic.edu/~liub/FBS/
sentiment-analysis.html
Collected Dataset Format Details
This dataset (HL-11prods-2200comments.xml) has
classification labels (from the manual annotation process
done by Dr. Bing Liu’s group) for the “opinion” class - which
marks whether or not a review comment consists of any
subjective evaluation of one or more features of the product
or the product itself [26]. The format of the file is pseudoXML. Each review comment is represented by an
<instance>...</instance> tag in the file. The complete set of
2,200 instances is enclosed in an outermost level
<instances>...</instances> tag. The <instance> tag has two
attributes - “id” and “subpop.” “id” is a unique identifier
given to each instance. “subpop” is a string that identifies
the product name for which the review comment was written.
Within each <instance> tag, the “cname” attribute in the
<class> tag contains the classification label - POS stands for
the opinion class, and NEG for the non-opinion class. The
<text>...</text> tag contains the actual text of the review
comment.
Data preprocessing
From this dataset format, we have been extracted data
from XML file and made it suitable for proposed design

6. Poster Paper
Proc. of Int. Conf. on Advances in Information Technology and Mobile Communication 2013
architecture in its data preprocessing step and applied feature
extraction module on it.
Feature Extraction
In this step, the use of text processing toolkit NLTK is
proposed along with Python to do natural language
processing for sentiment analysis. The results are in primary
stage.
[10] Hong Liu , “Internet public opinion hotspot detection and
analysis based on K-means and SVM algorithm,” International
Conference of Information Science and Management
Engineering, pp.257-261, 2010.
[11] Hui Wang, Jiansheng Chen, “Extracting Two-Noun phrases
from customer reviews,” IEEE 978-1-4244-4507-3, 2009.
[12] J. Wiebe, T. Wilson, R. Bruce, M. Bell, and M. Martin,
“Learning subjective language,” Computational Linguistics, vol.
30, pp. 277–308, 2004.
[13] Jianxiong Wang, Andy Dong, “A comparison of two text
representations for sentiment analysis,” in IEEE International
Conference on Computer Application and System Modeling
(ICCASM 2010), pp.35-39,2010.
[14] Khairullah Khan, Baharum B. Baharudin, “Identifying Product
Features from Customer Reviews using Lexical Concordance,”
Research Journal of Applied Sciences Engineering and
Technology 4(7), pp.833-839, 2012.
[15] Khairullah Khan, Baharum B.Baharudin, Aurangzeb Khan,
Fazal-e-Malik, “Mining opinion from text documents: a
survey,” 3rd IEEE International Conference on Digital
Ecosystems and Technologies, pp. 217-222, 2009.
[16] Luole Qi, Li Chen,”Comparison of Model-Based Learning
Methods for Feature-Level Opinion Mining,” in IEEE/WIC/
ACM International Conferences on Web Intelligence and
Intelligent Agent Technology, pp.265-273,2011.
[17] M. Hu and B. Liu , “Mining and Summarizing Customer
Reviews,” Proceedings of the ACM SIGKDD Conference on
Knowledge Discovery and Data Mining (KDD), pp. 168–
177,2004.
[18] Magdalini Eirinaki, Shamita Pisal , Japinder Singh,” Featurebased opinion mining and ranking,” Journal of Computer and
System Sciences ,pp.1175–1184, 2012.
[19] Mikalai Tsytsarau, Themis Palpanas, “Survey on mining
subjective data on the web,” In Data Min Knowl Disc , pp.
478–514,2012.
[20] Minqing Hu and Bing Liu ,”Mining opinion features in
customer reviews,” American Association for Artificial
Intelligence, pp.1-6, 2004.
[21] Rui Xia, Chengqing Zong, Shoushan Li, “Ensemble of feature
sets and classification algorithms for sentiment classification,”
ELSEVIER Information Sciences, pp. 1138–1152, 2011.
[22] Xiaowen Ding, Bing Liu, Philip S. Yu, “A holistic lexiconbased approach to opinion mining,” WSDM’08, ACM, pp.19, 2008.
[23] Yi Hu, Wenjie Li, “Document sentiment classification by
exploring description model of topical terms,” Science Direct
Computer Speech and Language, pp. 386–403, 2011.
[24] Yulan He, Deyu Zhou, “Self-training from labeled features for
sentiment analysis,” Information Processing and Management,
pp. 606–616, 2011.
[25] Zhixing Li, “Product feature extraction with a combined
approach,” IEEE Third International Symposium on Intelligent
Information Technology and Security Informatics, pp.686690, 2010.
[26] Joshi, Rose, “Opinion Mining Dataset,” ACL-IJCNLP 2009.
[27] Chandrashekhar D. Badgujar, “Opinion mining: extracting
and analyzing customers opinion on the Internet,” Proc.of
the second International Conference on Computer
Applications [ ICCA 2012] , pp. 29-33, 2012.
CONCLUSION
Finally, despite of the challenges, the opinion mining or
sentiment analysis has made significant progress over the
past few years. This is evident from the large number of startup companies that provide sentiment analysis or opinion
mining services. The opinions can be taken from all possible
means through media of newspaper, focus groups, blogs,
sms services and web sites also. Especially, it can be taken in
different languages other than English for the region wise or
country wise customer opinion reviews. So to enhance the
performance of sentiment or opinion analysis system, the
proposed design architecture of sentiment analysis will
integrate the multiple techniques of opinion analysis
altogether and will enrich the trust of people on such
technology.
REFERENCES
[1] Arjun Mukherjee, Bing Liu, “Aspect extraction through SemiSupervised modeling,” In support National Science Foundation
under grant no. IIS-1111092, pp.1 - 10, 2012.
[2] B. Liu , Sentiment Analysis and Subjectivity Handbook of
Natural Language Processing, Second Edition, 2010.
[3] Bing Liu, “Sentiment analysis: a multi-faceted problem,” IEEE
Intelligent Systems, pp.1-5, 2010.
[4] Bing Liu, Minqing Hu, Junsheng Cheng, “Opinion observer:
analyzing and comparing opinions on the web,” International
World Wide Web Conference Committee (IW3C2), ACM,
pp.1-10 , 2005.
[5] B. Pang and L. Lee, “Opinion mining and sentiment analysis,”
Foundations and Trends in Information Retrieval 2 (1-2), pp.
1–135, 2008.
[6] Chee Kian Leong, Yew Haur Lee, Wai Keong Mak, “Mining
sentiments in SMS texts for teaching evaluation,” Expert
Systems with Applications, pp. 2584–2589, 2012.
[7] Chunxia Yin, Qinke Peng, “Sentiment Analysis for Product
Features in Chinese Reviews Based on Semantic Association,”
International Conference on Artificial Intelligence and
Computational Intelligence, pp.82-85, 2009.
[8] Fazel Keshtkar, Diana Inkpen, “Using sentiment orientation
features for mood classification in blogs,” IEEE 978-1-42444538-7, pp. 1-6, 2009.
[9] Hanxiao Shi, Guodong Zhou, Peide Qian , “An attribute based
sentiment analysis system,” Information Technology Journal
ISSN 1812-5638, pp.1607-1614, 2010.
© 2013 ACEEE
DOI: 03.LSCS.2013.2.79
53