Docsity
Docsity

Prepare for your exams
Prepare for your exams

Study with the several resources on Docsity


Earn points to download
Earn points to download

Earn points by helping other students or get them with a premium plan


Guidelines and tips
Guidelines and tips

Challenges & Applications of Sentiment Analysis in Social Media, Summaries of Network Design

An overview of Sentiment Analysis (SA) and Opinion Mining (OM) techniques used in social media. The author, Ms. Latha S S, discusses the process of SA, its applications, and the challenges faced in implementing these techniques. SA is used to extract opinions and emotions from text data, which can be beneficial for businesses and consumers in various ways. the machine learning approach, specifically Naive Bayes and Support Vector Machine, and the lexicon-based approach for sentiment analysis. The applications of sentiment analysis include buying products or services, quality improvement, marketing research, opinion spam detection, policy making, and decision making.

What you will learn

  • What is the process of Sentiment Analysis and Opinion Mining?
  • How can Sentiment Analysis be used in marketing research?
  • What are the main challenges in implementing Sentiment Analysis techniques?

Typology: Summaries

2020/2021

Uploaded on 04/27/2022

ashww
ashww 🇮🇳

3 documents

1 / 6

Toggle sidebar

This page cannot be seen from the preview

Don't miss anything!

bg1
10 | P a g e
AN OPINION MINING AND SENTIMENT ANALYSIS
TECHNIQUES INVOLVING SOCIAL MEDIA:
CHALLENGES AND APPLICATIONS
Ms. Latha S S,
Assistant Professor, Department of Information Science & Engineering,
MVJ College of Engineering, Bengaluru, INDIA
ABSTRACT:
Sentiment Analysis (SA) is an on-going field of research in text mining field. Sentiment Analysis is a Process of
finding out extracting experiences and emotions from the given dataset. The two expressions Sentiment Analysis
or opinion mining are interchangeable. They express a mutual meaning.By using sentiment analysis on the
reviews of the customer in an e-commerce websites and enterprises can place a major change in the decision
making process.There are different procedures while making a sentiment analyser, Data Collection, data pre-
processing,Frequency computation,feature extraction and training with an algorithm are some of the steps
involved in the methodology. The main target of this survey is to give nearly full image of various challenges
while making a sentiment analyser and we are going to survey different techniques on sentiment analysis. Naive
Bayes and Support Vector Machine are the mostly used classifiers Further we discuss various challenges in
sentiment analysis.
Keywords: Lexicon based approach, Naive Bayes, Opinion Mining, Sentiment Analysis, Sentiment
Analyser and Support Vector Machine.
I. INTRODUCTION
Businesses and consumers buying and selling products in online refers to the E-commerce. The most of the e-
commerce websites sell products to the public directly. A review refers to the evaluation of a service,
publication, review of movies, video game review, review of a music composition or music recording, book
review, hardware piece like a car or computer, performance of a event, such as a live music concert, play,
musical theater show, dance show, or exhibition of a art .People often take reviews from their friends or
relatives who have bought the product before buying it. In today’s time reviews and ratings of the products plays
major role to generate opinion. To handle these problems Sentiment Analysis is used. Emotions of a sentence
can easily understand using sentiment analysis.
Data mining is an integrative part of computer science. It is the gauge process of searching patterns in data sets
involving methods at the junction of artificial intelligence, statistics, machine learning, and database systems.
The total goal of the sentiment analysis process is to extract information from a data set and convert it into a
suitable structure for further use. Opinion mining or sentiment analysis is to extract and classify the people’s
pf3
pf4
pf5

Partial preview of the text

Download Challenges & Applications of Sentiment Analysis in Social Media and more Summaries Network Design in PDF only on Docsity!

AN OPINION MINING AND SENTIMENT ANALYSIS

TECHNIQUES INVOLVING SOCIAL MEDIA:

CHALLENGES AND APPLICATIONS

Ms. Latha S S,

Assistant Professor, Department of Information Science & Engineering,

MVJ College of Engineering, Bengaluru, INDIA

ABSTRACT:

Sentiment Analysis (SA) is an on-going field of research in text mining field. Sentiment Analysis is a Process of finding out extracting experiences and emotions from the given dataset. The two expressions Sentiment Analysis or opinion mining are interchangeable. They express a mutual meaning.By using sentiment analysis on the reviews of the customer in an e-commerce websites and enterprises can place a major change in the decision making process.There are different procedures while making a sentiment analyser, Data Collection, data pre- processing,Frequency computation,feature extraction and training with an algorithm are some of the steps involved in the methodology. The main target of this survey is to give nearly full image of various challenges while making a sentiment analyser and we are going to survey different techniques on sentiment analysis. Naive Bayes and Support Vector Machine are the mostly used classifiers Further we discuss various challenges in sentiment analysis.

Keywords : Lexicon based approach, Naive Bayes, Opinion Mining, Sentiment Analysis, Sentiment

Analyser and Support Vector Machine.

I. INTRODUCTION

Businesses and consumers buying and selling products in online refers to the E-commerce. The most of the e- commerce websites sell products to the public directly. A review refers to the evaluation of a service, publication, review of movies, video game review, review of a music composition or music recording, book review, hardware piece like a car or computer, performance of a event, such as a live music concert, play, musical theater show, dance show, or exhibition of a art .People often take reviews from their friends or relatives who have bought the product before buying it. In today’s time reviews and ratings of the products plays major role to generate opinion. To handle these problems Sentiment Analysis is used. Emotions of a sentence can easily understand using sentiment analysis.

Data mining is an integrative part of computer science. It is the gauge process of searching patterns in data sets involving methods at the junction of artificial intelligence, statistics, machine learning, and database systems. The total goal of the sentiment analysis process is to extract information from a data set and convert it into a suitable structure for further use. Opinion mining or sentiment analysis is to extract and classify the people’s

opinion automatically from the internet. Sentiment Analysis uses a Natural Language Processing technique to identify positive, negative or neutral comments. To be specific, in a given piece of text, opinion mining aims to identify the part which is expressing the opinion and what is being communicated.

Figure 1 Sentiment analysis process on product reviews.

Sentiment Analysis is also considered a classification process as illustrated in Figure 1. First collect the reviews of products from the web and then parse the reviews to clean collected information. Cleaned data are divided to determine tokens. Once the token is identified it computes the frequency of identified keywords. Thefrequencies of keywords are used to represent features in our proposed model. The FEM matrix is constructed by using the list of Features to find the rank of product.

II METHODOLOGY

There are a number of techniques available for analysing and classifying sentiments to understand the opinions posted by individuals. There are three main approaches for analysing sentiments, namely [1]

2.1. Machine learning Approach : it applies machine learning algorithms with linguistic features and

can be implemented using either supervised learning or unsupervised learning methods [2]. It uses different types of algorithm to carry out the sentiment analysis. It includes training the particular portion of dataset and then using the remaining portion of dataset to test for the result. Majorly used algorithm is: ● Naïve Bayes Naïve Bayes algorithm is derived from Bayes' theorem. It consists of a family of algorithms. Bayes’ theorem computes the probability of given set using already calculated probabilities[3]. Figure 5 describes Bayes' Theorem mathematically

  1. Quality Improvement in Product or service: By sentiment analysis the manufactures cancollect the critic’s opinion as well as the favourable opinion about their product or service so they canimprove the quality of their product or service. They can utilize online product reviews from websites such asAmazon and RottenTomatoes.com.
  2. Marketing research: The result of sentiment analysis techniques can be utilized in marketing research [5]. Bysentiment analysis techniques, the recent trend of consumers about some product or services can be analysed. Modern attitude of general public towards some new government policy can also be easily analysed.These all result can be contributed to collective intelligent research.
  3. Recommendation Systems: By classifying the people’s opinion into positive and negative, the system can saywhich one should get recommended and which one should not get recommended [6].
  4. Detection of flame: Keep tracking of newsgroup and forums, blogs and social media is easily possible bysentiment analysis. Opinion mining and sentiment analysis can automatically detect arrogant words,overheated words or hatred language used in emails or forum entries or tweets on various internet sources.
  5. Opinion spam detection: Since internet is available to all, anyone can put anything on internet, this increased thepossibility of spam content on the web. People may write spam content to mislead the people. Opinion mining andsentiment analysis can classify the internet content into “spam” content and “not spam” content.
  6. Policy Making: Through Sentiment analysis, law makers can take citizen’s point of view towards some policyand they can utilize this information in creating new citizen friendly policy.
  7. Decision Making: People’s reviews and experience are very helpful element in decision making process. Opinion mining and Sentiment analysis gives analysedpeople’s opinion that can be effectively used for decision making.

IV RESEARCH CHALLENGES IN OPINION MINING AND SENTIMENT ANALYSIS

  1. Detection of spam and fake reviews: The web contains both authentic and spam contents. For effective Sentiment classification, this spam content should be removed before processing. This can be done by identifying duplicates, by detecting outliers and by considering reputation of reviewer.
  2. Multiple Language Input As the dataset is a collection of the reviews given by the users, it can be in multiple languages. But the classifier mainly uses English language. Therefore it becomes very difficult to train the algorithm for different languages other than English. Hence Multiple Language Input is a big challenge in sentiment analysis [7].
  3. Limitation of classification filtering: There is a restriction in classification filtering while determining most popular thought or concept. For finer sentiment classification result this limitation should be reduced. The risk of filter bubble gives immaterial opinion sets and it results false summarization of sentiment.
  4. Asymmetry in accessibility of opinion mining software: The opinion mining software is very expensive and currently affordable only to big organizations and government. It is beyond the common citizen’s expectation. This should be available to all people, so that everyone gets advantage from it.
  1. Incorporation of opinion with implicit and behaviour data: For successful analysis of sentiment, opinion words should integrate with implied data. The implied data determine the actual behaviour of sentiment words.
  2. Domain-independence: The major difficulty faced by opinion mining and sentiment analysis is the domain dependent nature of sentiment words. One features set may give excellent results in one domain, at the same time it perform very poor in some other domain.
  3. Emoticons and Sarcastic Reviews Emoticons are the pictorial representation of one's expressions. Using emoticons to define the product makes it easier for the customer or user to understand one's feelings. On the other hand it becomes difficult for a machine to understand the emoticons. It is not easy to train an algorithm for emoticons as an input. Sarcastic reviews are difficult to interpret by the machine. The model needs to be trained with more and more such data to give an accurate answer. Hence, Emoticons and sarcastic reviews are one of the biggest challenges of sentiment analysis [8].
  4. Natural language processing overheads: The natural language overhead like ambiguity, co-reference, inference etc. created hindrance in sentiment analysis too.

V CONCLUSIONS

Thus Sentiment analysis has wide area of applications and it also facing many research challenges. Since the fast growth of internet and internet related applications, the Opinion Mining and Sentiment Analysis become a most interesting research area among natural language processing community. A more innovative and effective techniques required to be invented which should overcome the current challenges faced by Opinion Mining and Sentiment Analysis.

REFERENCES

[1]Diana Maynard, Adam Funk. Automatic detection of political opinions in tweets. In: Proceedings of the 8th

international conference on the semantic web, ESWC’11; 2011.

[2] Xu, Yun, Xinhui Wu, and Qinxia Wang. "Sentiment Analysis of Yelp„s Ratings Based on Text Reviews."

[3] Latha S S,” AN EXPERIMENTAL ANALYSIS ON ECOMMERCE REVIEWS, WITH SENTIMENT

CLASSIFICATION USING OPINION MINING ON WEB”,IJEAST- , Vol. 5, Issue 11, ISSN No. 2455-2143, Pages 274-

[4] Liu, B. (2010), “Sentiment Analysis and Subjectivity”. Appeared in Handbook of Natural Language

Processing, Indurkhya, N. &Damerau, F.J. [Eds.].

[5] Boiy, E., Hens, P., Deschacht, K. &Moens, M.-F. (2007), “Automatic Sentiment Analysis in OnLine Text”.

In Proceedings of the Conference on Electronic Publishing (ELPUB-2007), p. 349-360.

[6] Latha.S.S,” Analysing the reviews and comments of customers in an e-commerce websites”, JEST-M May-

2017 Issue-5 Page no[1-4],ISSN:: 2394 -