help

Sentiment_Analysis_of_Twitter_Data_in_Online_Social_Network.pdf

Home >Computer Science homework help >help

5 th IEEE International Conference on Signal Processing, Computing and Control (ISPCC 2k19), Oct 10-12, 2019, JUIT, Solan, India

Sentiment Analysis of Twitter Data in Online Social Network

Sanjeev Dhawan Department of Computer Science and

Engineering University Institute of Engineering and

Technology(UIET),Kurukshetra University,

Kurukshetra-136119 Kurukshetra,India

[email protected]

Kulvinder Singh Department of Computer Science and

Engineering University Institute of Engineering and

Technology(UIET),Kurukshetra University,

Kurukshetra-136119 Kurukshetra,India

[email protected]

Priyanka Chauhan Department of Computer Science and

Engineering

University Institute of Engineering and Technology(UIET),Kurukshetra University,

Kurukshetra-136119

Kurukshetra,India [email protected]

Abstract- Sentiment Analysis is the procedure of computationally

deciding if a bit of composing is certain, negative or nonpartisan.

It's otherwise called supposition mining, inferring the sentiment

or frame of mind of a user. In this paper, an attempt has been

made to propose analysis method for sentiment of twitter

dataset. In proposed method polarity of each tweet is calculate to

distinguish whether tweet is positive or negative. A sentiment

polarity is the emotions of user such as angry, sad, happy and

joy. The proposed mechanism has been implemented in Python.

Keywords— Sentiment Analysis, Twitter, Sentiment Polarity,

Supervised Learning, Unsupervised Learning.

I. INTRODUCTION

Sentiment is naturally passionate from individual to

individual, and can even be absolute senseless. It's essential to

mine a gigantic and relevant case of data when attempting to

evaluate sentiment. No particular data point is basically

relevant. An individual's sentiment toward a brand or thing

may be influenced by in any event one indirect causes;

someone may have a horrendous day and tweet an unfriendly

remark about something they for the most part had a really

unprejudiced supposition about. With a huge enough

example, exceptions are weakened in the aggregate [1].

Additionally, since sentiment likely changes after some time

as indicated by an individual's state of mind, world occasions,

etc to investigations the sentiments of clients on twitter [2].

Albeit various analysts proposed various methods to analysis

sentiments of clients on twitter informational collection. The

current strategies works just on the dataset which is

compelled to a specific point. The current procedures likewise

don't decide the proportion of effect the outcomes decided can

have on the specific field mulled over and it doesn't permit

recovery of information dependent on the inquiry entered by

the client for example it has obliged scope. In straightforward

words, it chips away at static information as opposed to

dynamic data [3]. Unsupervised calculations like Vector

Quantization are utilized for information pressure, design

acknowledgment, facial and discourse acknowledgment, and

so forth and in this manner can't be utilized in deciding

sentiment in twitter information. Apriori calculation neglects

to deal with enormous datasets and accordingly can produce

flawed results [4-5]. This paper is divided into IV sections.

Section I presents introduction while section II covers

background work, in section III proposed work has been

presented at last in section IV conclusion of paper has been

discussed.

II. BACKGROUND WORK

There are for the most part two procedures for sentiment

analysis for the twitter information:

AI Approaches

AI based methodology utilizes grouping strategy to arrange

content into classes. There are for the most part two kinds of

AI techniques [6] [7].

Authorized licensed use limited to: University of the Cumberlands. Downloaded on October 17,2021 at 14:56:41 UTC from IEEE Xplore. Restrictions apply.

5 th IEEE International Conference on Signal Processing, Computing and Control (ISPCC 2k19), Oct 10-12, 2019, JUIT, Solan, India

•Unsupervised learning:

It doesn't comprise of a classification and they don't give the

right focuses at all and in this manner depend on bunching.

•Supervised learning:

It depends on named dataset and in this way the marks are

given to the model during the procedure. These named dataset

are prepared to get significant yields when experienced during

choice making [8].

The achievement of both this learning strategies is primarily

relies upon the determination and extraction of the particular

arrangement of highlights used to distinguish sentiment. The

AI approach relevant to sentiment analysis primarily has a

place with supervised classification [9]. In an AI method, two

arrangements of information are required:

•Training Set

•Test Set.

AI begins with gathering preparing dataset. Nextly we train a

classifier on the preparation data[10]. When a supervised

characterization system is chosen, a significant choice to

make is to choose include. They can disclose to us how

archives are spoken to.

The most regularly utilized highlights in sentiment grouping

are

•Term nearness and their recurrence

•Part of discourse data

•Negations

III. PROPOSED WORK

In promoting field organizations, use it to build up their

methodologies, to comprehend clients' sentiments towards

items or brand, how individuals react to their battles or item

dispatches and why buyers don't get a few items. In political

field, it is utilized to monitor political view, to identify

consistency and irregularity among proclamations and

activities at the administration level. It tends to be utilized to

foresee decision results also. Sentiment analysis likewise is

utilized to screen and examinations social marvels, for the

spotting of possibly perilous circumstances and deciding the

general state of mind of the blogosphere.

Proposed Algorithm

1. Start

2. Generate dataset from twitter.

3. Fetch secret keys and secret key tokens from twitter

dataset.

4. Check authenticity of user.

5. If (secret key ≠ original secret key_ twitter user)

{

Then display authentication failed.

Else

Check sentiment polarity of each tweet.

(For i=1; i<n; i++)

{

Extract different sentiments of users.

If (sentiment polarity > 0)

{

Print positive tweet.

If (sentiment polarity = 0)

Authorized licensed use limited to: University of the Cumberlands. Downloaded on October 17,2021 at 14:56:41 UTC from IEEE Xplore. Restrictions apply.

5 th IEEE International Conference on Signal Processing, Computing and Control (ISPCC 2k19), Oct 10-12, 2019, JUIT, Solan, India

{

Print neutral tweet.

Else

Negative tweet.

}

6. End.

Description of proposed algorithm

In the proposed algorithm, twitter dataset is fetched from

twitter API for analysis of sentiments emotions of different

users. Here first check authenticity of user with the help of

secret key of user and secret key tokens. If the authenticity

of users failed then display a message that the user is

unauthorized. If user is unauthorized then algorithms stops to

perform sentimental analysis otherwise start sentimental

analysis. To analyze sentiments first extract features of

twitters to analyze sentiments of twitter. Here we check

sentiment polarity of each tweet. The sentiment polarity is

emotions of users like joy, happy, sad and angry. If the

sentiment polarity is equal to zero then tweet is neutral and if

polarity is greater than zero then tweet is positive otherwise,

tweet is negative. In this manner proposed algorithm

distinguish tweets based on sentiment polarity of each tweet

of users.

IV. RESULTS

To implement proposed method python language is used. It is

a open source language and run on any platform.

Fig 2 distinguishment of tweets

(Fig 2) depicts distinguishment of tweets in two categories i.e.

positive tweets, negative tweets and neutral tweets

Authorized licensed use limited to: University of the Cumberlands. Downloaded on October 17,2021 at 14:56:41 UTC from IEEE Xplore. Restrictions apply.

5 th IEEE International Conference on Signal Processing, Computing and Control (ISPCC 2k19), Oct 10-12, 2019, JUIT, Solan, India

Fig 3 positive tweets and negative tweets

(Fig 3) depicts actual positive tweets with their messages and

negative tweets with their messages.

V. Conclusion

In this paper prose a sentiment polarity based sentiments

analysis mechanism for twitter dataset in online social

networks. In the proposed algorithm, twitter dataset is

fetched from twitter API for analysis of sentiments

emotions of different users. Here we check sentiment

polarity of each tweet. The sentiment polarity is

emotions of users like joy, happy, sad and angry. If the

sentiment polarity is equal to zero then tweet is neutral

and if polarity is greater than zero then tweet is positive

otherwise, tweet is negative. In this manner proposed

algorithm distinguish tweets based on sentiment polarity

of each tweet of users.

Future work: In future it is intended to continue working

on it and propose a new sentiments analysis method for

analyzing image and video sentiments of twitter data.

REFERENCES

[1] C. N. Dos Santos and M. Gatti, Deep convolutionalneural networks for

sentiment analysis of short texts, in Proc. COLING 2014, the 25th Int. Conf. Computational Linguistics: Technical Paper, Dublin, Ireland, 2014, pp. 69–78.

[2] M. Hasan, E. Rundensteiner, and E. Agu, EMOTEX: Detecting emotions in twitter messages, in Proc. 2014ASE BIGDATA/SOCIALCOM/CYBERSECURITY Conf.,Stanford University, CA, USA, 2014.

[3] S. J. Liu, N. Yang, M. Li, and M. Zhou, A recursive recurrent neural network for statistical machine translation, in Proc. 52nd Annual Meeting of the Association for Computational Linguistics, Baltimore, MD, USA, 2014, pp. 1491–1500.

[4] E. S. Tellez, S. Miranda-Jim´enez, M. Graff, D. Moctezuma, O. S. Siordia, and E. A. Villase˜nor, A case study of Spanish text transformations for twitter sentiment analysis, Expert Syst. Appl., vol. 81, pp. 457–471, 2017.

[5] M. Bouazizi and T. Ohtsuki, Sentiment analysis: From binary to multi- class classification: A pattern-based approach for multi-class sentiment analysis in twitter, in Proc. 2016 IEEE Int. Conf. Communications, Kuala Lumpur, Malaysia, 2016, pp. 1–6.

[6] M. Fujita, J. I. Watanabe, K. Kawamoto, T. Akitomi, andK. Ara, A method for analyzing influence of emotionsof posts in SNS conversations, in Proc. 2013 Int. Conf.Social Intelligence and Technology, State College, PA, USA, 2013, pp. 20–27.

[7] X. Z. Ruan, S. R. Wilson, and R. Mihalcea, Finding optimists and pessimists on twitter, in Proc. 54th Annual Meeting of the Association for Computational Linguistics, Berlin, Germany, 2016, pp. 320–325.

[8] A. Taketomi and M. Hisano, Twitter users’ characteristics and their emotional expressions in the tweets, IEICE Technical Report, NLC2014-34, vol. 114, No. 366, pp. 1– 4, 2014. (in Japanese)

[9] H. Kwak, C. Lee, H. Park, and S. Moon, What is twitter, a social network or a news media?, in Proc. 19th Int. Conf. World Wide Web, Raleigh, North Carolina, USA, 2010, pp. 591–600.

[10] A. Java, X. D. Song, T. Finin, and B. Tseng, Why we twitter: Understanding microblogging usage and communities, in Proc. 9th WebKDD and 1st SNA-KDD 2007 Workshop on Web Mining and Social Network Analysis, San Jose, CA, USA, 2007, pp. 56–65.

[11] A. Tanaka, H. Takemura, and K. Tajima, Why you follow: A classification scheme for twitter follow links, in Proc. 25th ACM Conf. Hypertext and Social Media, Santiago, Chile, 2014, pp. 324–326.

[12] A. Koide, K. Saito, K. Kazama, and F. Toriumi, Study of twitter’s follow mechanism based on network analysis, (in Japanese),

Authorized licensed use limited to: University of the Cumberlands. Downloaded on October 17,2021 at 14:56:41 UTC from IEEE Xplore. Restrictions apply.

5 th IEEE International Conference on Signal Processing, Computing and Control (ISPCC 2k19), Oct 10-12, 2019, JUIT, Solan, India

Transactions on Mathematical Modeling and Its Applications, vol. 6, no. 2, pp. 164–173, 2013.

[13] M. Sato, M. Yoshikai, and S. Oguri, Study on analysis of community in twitter, (in Japanese), Forum Inf. Technol., vol. 12, no. 4, pp. 99–103, 2013.

Authorized licensed use limited to: University of the Cumberlands. Downloaded on October 17,2021 at 14:56:41 UTC from IEEE Xplore. Restrictions apply.