Prediction of 2019 Indian Election Using Sentiment Analysis
Bharat R.Naiknaware1st Dept. of Computer Science and IT,
Dr. Babasaheb Ambedkar Marathwada University, Aurangabad, India
Seema S.Kawathekar2nd Dept. of Computer Science and IT Dr. Babasaheb Ambedkar Marathwada University
Aurangabad, India [email protected]
Abstract— the aim of this paper is to study and compare people’s opinions on Indian Government Schemes to evaluate the reputation of Schemes using sentiment analysis. In this research paper developed on social media twitter datasets of particular schemes and its polarity of sentiments. Popularity of Internet has been rapidly increased. Sentiment analysis and opinion mining is the field of study that analyses people's opinions, sentiments, evaluations and attitudes, from written English language. The proposed research classify recent opinions into different categories based on Twitter datasets, positive polarity of schemes, and Negative polarity of schemes and Neutral Polarity of Schemes. In this research worked on Sentiment Analysis for the purpose of prediction of 2019 Indian election. For that proposed work used sentiment analysis score technique. Sentiment Analysis is the computational evaluations of opinions, feedbacks, discussions, emotions present in the text. From text to find the meaningful sentence polarity is important task. Sentiment analysis over Twitter offers people a fast and effective way to measure the public’s feelings towards their party and politicians.
Here we have eight major schemes and their last three years twitter datasets from year 2016 to year 2018. All the twitter datasets are created by own during research work. On the basis of that data and its polarity. Found meaningful patterns and we get easy prediction to 2019 Indian Election. We were used this sentiment classifier for finding political sentiment from real time tweets. In the field of politics, where political entities need to understand public opinion and thus determine their campaigning strategy. Sentiment analysis on social media data has been seen by many as an effective tool to monitor user preferences and inclination. For this eight schemes for prediction that are GST, Demonetization, Digital India, Make In India, Startup India, Swacha Bharat, Kashmir, Yoga Day with minimum 1500 tweets are compared. Finally for prediction purpose we focused on Positive and Negative polarity of datasets and this is base for prediction of 2019 Indian election.
Keywords—Score, Sentiment Analysis, Twitter API, GST, ROAuthm,JSON.
I. INTRODUCTION
Social media is an internet-based form of communication.
Social media platforms allow users to have discussions, share information and create web content. There are many forms of social media, including blogs, micro-blogs, wikis, social networking sites, photo-sharing sites, instant messaging, video-
sharing sites, podcasts, widgets, virtual worlds. In the social media world the web content has become the main attraction. It gives useful information and entertainment to Internet users or a business opportunity for marketing companies and content providers. Web content is a valuable asset on the Internet. In the world the social media is increasing rapidly with new innovations conversion of data to information [1]. The Twitter data is particularly appropriate for our study since Twitter is the widely used and popular micro blogging service in the election prediction as well sharing opinion regarding governments work and we are mainly interested in the government scheme for people’s opinion. Consequently, the analysis of Twitter data can provide the insights into opinions and discussions about governments. In order to analyze opinions in tweets, we apply sentiment analysis, which is the research area aiming at detecting the people’s attitude, emotions, or opinion about a given topic expressed in text where the word sentiment represents an attitude, view, or opinion caused by emotion. It is important to perform appropriate preprocessing of such data in order to prepare it in the best possible way as input of sentiment analysis algorithms [4]. Basic task in sentiment analysis is to classify text as being positive, negative, or neutral. While some approaches perform more complex analyses. Recently, there has been an increasing trend and adoption of social media by government organizations for not just disseminating information but also collecting information such as discussion on current decisions, complaints and grievances from peoples. In particular, social media platforms like twitter and Facebook are gaining popularity as social-media based grievance management system or platforms on which peoples can lodge complaints.
Twitter is one of the most widely used micro-blogging websites on Internet. Due to the wide reachability and connectivity among its users, Twitter is being used by the National Government as well state governments to reach out the public.
II. SURVEY OF LITRATURE
In this survey of literature here provide recent work on sentiment analysis and opinion mining related to my work.
Aliza Sarlan1, Chayanit1 Nadam2, Shuib Basri3 they works on Twitter Sentiment Analysis Social media has received more attention nowadays. Public and private opinion about a wide
variety of subjects are expressed and spread continually via numerous social media. Twitter is one of the social media that is gaining popularity this paper reports on the design of a sentiment analysis, extracting a vast amount of tweets.
Prototyping is used in this development. Results classify customers’ perspective via tweets into positive, negative and neutral. To associate with Twitter API, developer need to agree in terms and conditions of development Twitter platform which has been provided to get an authorization to access a data. The output from this process will be saved in JSON file. The reason is, JSON (JavaScript Object Notation) is a lightweight data- interchange format which is easy for humans to write and read [10].
III. METHODOLOGY
Winner of any election is decided on who is being voted for in majority. But opinions are not formed in one day. By observing the political opinion of a person, one can predict which party the person would vote for. Twitter provides a platform for people to their opinion in a precise manner, i.e. in just 280 characters. If tweets are stored gradually a year, we could analyze what the opinion of the people and predict who is going to win the Election. This paper focuses on seven main things:
1) Extracting relevant tweets using Twitter API 2) Tweets Preprocessing
3) Storing the processed tweets in CSV File format 4) Apply Score.Sentiment() method
5) Generate Sentence Score
6) Analyzing the sentiment polarity of each tweet.
7) Prepare results.
Figure 1. Methodology Score.Sentiment ()
A. Data Collection:
Determining the data source is the first challenge of any research. Based on debates and news, the sentiments related to elections are changing rapidly. In order to handle this temporal nature data, Twitter API is used for streaming tweets and a
Library called twitteR which enables R to use this API. we are collected the Eight hashtag tweets from twitter minimum 5000 Tweets of GST, Demonetization ,Digital India, Make in India, Swacha Bharat, Startup India, Kashmir and Yoga day.
Standard twitter dataset is not available for Election related data domain, we created a new dataset by collecting tweets over a period of 2016 to 30/06/2018 day time data. Tweets are collected using Twitter API with Twitter Application Development and they are manually annotated as positive or negative. We are creating Twitter Development Application account after that twitter provide access key, Secret Key, access Token and Application Authentication ID this credential used for fetching data from Twitter Account. We are using following code for extracting the dataset from twitter API.
Library (twitteR) Library (ROAuth) Library (plyr)
CUSTOMER_KEY <- "..."
CUSTOMER_SECRET <-"..."
ACCESS_TOKEN <- "..."
ACCESS_secret <- "..."
setup_twitter_oauth(CUSTOMER_KEY,
CUSTOMER_SECRET, ACCESS_TOKEN, ACCESS_secret) 1
GST2018 = searchTwitter("#GST2018 ", n= 5000, lang="en",since='2018-04-15', until='2018-05-12')
GST2018tweets_df = twListToDF(GST) write.csv(tweets_df,
file='C:/Users/BHARAT/Desktop/DataSet/ GST2018.csv', row.names=F).
B. Data Preprocessing:
After extracting the tweets, we then had to determine how to extract the sentiments from these tweets. The tweet input text is processed and then converted all of the text to lowercase, removed hashtags before words, and removed extra spaces in the text.[3] Since Twitter is very casual, people often include multiples of the same letters in a word, such as “happyyyyy, awesommm.” To handle these cases, more than two of the same consecutive letters are replaced with just two of that letter. All URLs and usernames are ignored since URLs and usernames don’t typically relate to the sentiments expressed in the tweet. Furthermore, a list of common words that do not express sentiment words like “be,” “at,” “the,” etc are also ignored. Keyword extraction is difficult in twitter due to misspellings and slang words. So to avoid this, a preprocessing step is performed before feature extraction. Preprocessing steps include removing URL, avoiding misspellings and slang words. Misspellings are avoided by replacing repeated characters with 2 occurrences. Slang words contribute much to the emotion of a tweet, Remove RT, Remove Hashtags, Remove Controls and special characters, Remove Controls and
special characters, Remove Punctuations, Remove leading whitespaces, Remove trailing whitespaces, Remove extra whitespaces . So they can’t be simply removed. Therefore a slang word dictionary is maintained to replace slang words occurring in tweets with their associated meanings. Domain information contributes much to the formation of slang word dictionary.
Library (stringr) Library(tm)
Dataset <-
read.csv('C:/Users/BHARAT/Desktop/DataSet/Paper/
GST2018.csv')
Dataset$text <- as.factor(Dataset$text)
pos.words <-
scan('C:/Users/BHARAT/Desktop/DataSet/Paper/pos_words.tx t', what='character', comment.char=';') #folder with positive dictionary
neg.words <-
scan('C:/Users/BHARAT/Desktop/DataSet/Paper/neg_words.t xt', what='character', comment.char=';') #folder with negative dictionary
#pos.words <- c(pos, 'upgrade')
#neg.words <- c(neg, 'wtf', 'wait', 'waiting', 'epicfail')
score.sentiment <- function(sentences, pos.words, neg.words, .progress='none')
{
require(plyr) require(stringr)
scores <- laply(sentences, function(sentence, pos.words, neg.words){
sentence = gsub("(RT|via)((?:\\b\\W*@\\w+)+)", "", sentence) sentence <- gsub('[[:punct:]]', "", sentence)
sentence <- gsub('[[:cntrl:]]', "", sentence) sentence <- gsub('\\d+', "", sentence
sentence = gsub('(RT|via)((?:\\b\\W*@\\w+)+)', '', sentence) sentence <- tolower(sentence)
word.list <- str_split(sentence, '\\s+') words <- unlist(word.list)
pos.matches <- match(words, pos.words) neg.matches <- match(words, neg.words) pos.matches <- !is.na(pos.matches) neg.matches <- !is.na(neg.matches)
score <- sum(pos.matches) - sum(neg.matches) return(score)
}, pos.words, neg.words, .progress=.progress)
scores.df <- data.frame(score=scores, text=sentences) return(scores.df)
}
scores<-score.sentiment(Dataset$text, pos.words, neg.words, .progress='text')
write.csv(scores,file=('C:/Users/BHARAT/Desktop/DataSet/Pa per/Score GST2018.csv'), row.names=TRUE
C. Generate Score for sentiment Analysis:
The most important part of sentiment analysis to generate score each tweet, score. Sentiment ( ) function is used to iterate through the input text. It strips punctuation and control characters from each line using in R Programming platform
regular expression-powered substitution function, and matches against each word list to find matches. The score. Sentiment ( ) function assigns score to the tweets using the formula as Score
= sum (pos.matches) – sum(neg.matches) The score is maintained between -5 to 5. 5 and 4 represent very positive -5 and -4 represent very negative 2 and 1 represent positive -2 and -1 represent negative if the score turns out to be zero, it is classified as neutral.
IV. RESULTS INTERPREATION
Here we got meaningful sentiment from different eight datasets and we are predicted the 2019 Indian Election based on score.
Sentiment ( ) function the approach followed here is to count the positive and negative words in each tweet and assign a sentiment score. This way, we can ascertain how positive or negative a tweet is. There are multiple ways to calculate such scores here is one formula to perform such calculations.
Score = Number of positive words - Number of negative words
If Score > 0, means that the tweet has 'Positive Sentiment' If Score < 0, means that the tweet has 'Negative Sentiment' If Score = 0, means that the tweet has 'Neutral Sentiment' Results are computed in R Open Source tool here we are using twitteR, plyr, ROAuth, tm,NLP,stringr ggplot2 packages for calculating polarity Score. Comparative sentiment analysis eight major declared schemes by central government after 2014. Here we are predicted the people’s opinion on every individual scheme hashtag the own datasets are created and apply the sentiment analysis techniques. With help of sentiment analysis score generation methods we are generated score of every sentence tweets and score are classifiers into three types polarity feature. The most important part of sentiment analysis to generate score each tweet, score.
Sentiment ( ) function is used to iterate through the input text.
It strips punctuation and control characters from each line using in R Programming platform regular expression-powered substitution function, and matches against each word list to find matches.
1) Positive Polarity: Positive score defines the sentence having positive words is called as Positive Polarity.
In the Positive polarity the numbers are started from 1 to 5 meaning is in score generated sentence having maximum 5 words are positive or at least one word is positive.
2) Negative Polarity: Negative score defines the sentence having negative words is called Negative Polarity.
In the negative polarity the numbers are started from -1 to -5 meaning is in score generated sentence having maximum 5 words are negative or at least one word is negative.
3) Neutral Polarity: Neutral Polarity having score value is 0 means it indicates in the sentence there is no positive or negative words are presents the sentence generates the 0 score i.e. Neutral Polarity.
The score. Sentiment ( ) function assigns score to the tweets using the formula as Score = sum (pos.matches) – sum
(neg.matches). The score is maintained between -5 to -1. 5 and 1 represent very positive -5 and -1 represent very negative 5 and 1 represent positive -5 and -1 represent negative if the score turns out to be zero, it is classified as neutral.
Here we are use three own datasets for prediction of 2019 India Election popularity issues.
Table I Comparative Analysis of GST twitter dataset
Dataset % Pos % Neg % Neu
GST2016 53.89 2.2 43.9
GST2017 40.8 8.32 50.88
GST2018 37.24 10.46 52.3
Figure 2. Comparative Analysis of GST twitter dataset
in above Comparative Analysis of GST twitter dataset table in year 2016 53.89% of tweets are categorized as positive, and in year 2017 40.8 % of tweets are positive, in year 2018 37.24%
of tweets categorized as positive in 2016 2.2 % of tweets categorized as negative and in year 2017 8.32 % of tweets are Negative, in year 2018 10.46 % of tweets categorized as negative and in year 2016 43.9 % of tweets are categorized as neutral and in year 2017 50.88 % of tweets are categorized as neutral, in year 2018 52.3 % of tweets are categorized as Neutral means in the year 2016 peoples are positive for GST and in year 2017 less peoples are positive and in year 2018 more peoples are neutral in three year peoples are having less negative opinion in year 2018 52.3% peoples are neutral opinions. Overall opinions of peoples on GST scheme is Positives.
Table II Comparative Analysis of Demonetization twitter dataset
Dataset % Pos % Neg % Neu
Demonetization 2016 20.77 30.82 48.39 Demonetization 2017 21.78 27.14 51.08 Demonetization 2018 11.16 45.52 43.32
Figure 3. Comparative Analysis of Demonetization twitter dataset in above Comparative Analysis of Demonetization twitter dataset table in year 2016 20.77 % of tweets are categorized as positive, and in year 2017 21.78 % of tweets are positive and in
year 2018 11.16 tweets are categorized as positive in 2016 30.82 % of tweets categorized as negative and in year 2017 27.14 % of tweets are Negative ad in year 2018 45.52 % of tweets are categorized as negative and in year 2016 48.39 % of tweets are categorized as neutral and in year 2017 51.08 % of tweets are categorized as neutral, in year 2018 43.32 % of tweets categorized as Neutral means in the year 2016 and year 2017 less peoples are positive for Demonetization and in year 2016 and year 2018 more peoples are negative in three year peoples are having more Neutral opinion in year 2017 51%
peoples are neutral opinions. Overall opinions of peoples on Demonetization scheme is Negative.
Table III Comparative Analysis of Digital India twitter dataset
Dataset % Pos % Neg % Neu
Digital India 2016 24.84 8.24 67.04 Digital India 2017 41.46 12.04 46.26
Digital India 2018 43.56 9.7 46.74
Figure 4. Comparative Analysis of Digital India twitter dataset
in above Comparative Analysis of Digital India twitter dataset table in year 2016 24.84% of tweets are categorized as positive, and in year 2017 41.46 % of tweets are positive and year 2018 43.56% of tweets are categorized as Positive in 2016 8.24 % of tweets categorized as negative and in year 2017 12.06 % of tweets are Negative in year 2018 9.7 % of tweets are categorized as Negative and in year 2016 67.04 % of tweets are categorized as neutral and in year 2017 46.26 % of tweets are categorized as neutral and in year 2018 46.74 % of tweets are categorized as Neutral means there is no exact people’s opinions on positive or negative. Overall opinions of peoples on Digital India scheme is Positives.
Table IV Comparative Analysis of Make in India twitter dataset
Dataset % Pos % Neg % Neu
Make in India 2016 33.87 4.55 61.64 Make in India 2017 52.82 9.64 37.66 Make in India 2018 49.56 9.4 41.04
Figure 5. Comparative Analysis of Make in India twitter dataset
in above Comparative Analysis of Make in India twitter dataset table in year 2016 33.87 % of tweets are categorized as positive, and in year 2017 52.82 % of tweets are positive and in year 2018 49.56 % of tweets are categorized as Positive in 2016 4.55 % of tweets categorized as negative and in year 2017 9.64 % of tweets are Negative, in year 2018 9.4 % of tweets are categorized as Negative and in year 2016 61.64 % of tweets are categorized as neutral and in year 2017 37.66 % of tweets are categorized as neutral and in year 2018 41.04% of tweets are categorized as neutral means in the year 2016 and year 2017 more peoples are positive for Make in India and in year 2016 and year 2017 less peoples are negative in 2016 year peoples are having more Neutral opinion i.e. 61.64% in year 2017 37.66 % peoples are less neutral opinions. Overall opinions of peoples on Make in India scheme is Positives.
Table V Comparative Analysis of Startup India twitter dataset
Dataset % Pos % Neg % Neu
Startup India 2016 37.12 7.72 55.4 Startup India 2017 36.68 6.1 57.34 Startup India 2018 45.66 6.56 47.76
Figure 6. Comparative Analysis of Startup India twitter dataset
in above Comparative Analysis of Startup India twitter dataset table in year 2016 37.12 % of tweets are categorized as positive, and in year 2017 36.68 % of tweets are positive and in year 2018 45.66 % of tweets are categorized as Positive in 2016 7.72 % of tweets categorized as negative and in year 2017 6.1 % of tweets are Negative in year 2018 6.56 % of tweets are categorized as negative tweets and in year 2016 55.4 % of tweets are categorized as neutral and in year 2017 57.34 % of tweets are categorized as neutral and in year 2018 47.76 % of tweets are categorized as neutral means in the year 2016 and year 2018 more peoples are positive for Startup India and in year 2017 and year 2018 less peoples are negative in 2017 year peoples are having more Neutral opinion i.e. 57.34 % in year 2016 55.4 % peoples are less neutral opinions. Overall opinions of peoples on Startup India scheme is Positives.
Table VI Comparative Analysis of Swacha Bharat twitter dataset
Dataset % Pos % Neg % Neu
Swacha Bharat 2016 44.08 9.11 46.08 Swacha Bharat 2017 44.06 10.34 45.06 Swacha Bharat 2018 29.52 17.52 52.94
Figure 7. Comparative Analysis of Swacha Bharat twitter dataset
in above Comparative Analysis of Swacha Bharat twitter dataset table in year 2016 44.08 % of tweets are categorized as positive, and in year 2017 44.06 % of tweets are positive in year 2018 29.52% of tweets are categorized as positive in 2016 9.11 % of tweets categorized as negative and in year 2017 10.34 % of tweets are Negative, in year 2018 17.52 % of tweets are categorized as negative and in year 2016 46.08 % of tweets are categorized as neutral and in year 2017 45.06% of tweets are categorized as neutral and in year 2018 52.94 % of tweets are categorized as neutral means in the year 2016 and year 2017 more peoples are positive for Swacha Bharat and in year 2016 and year 2017 less peoples are negative in 2016 year peoples are having more Neutral opinion i.e. 46.08 % in year 2017 45.06 % peoples are less neutral opinions in year 2018 more peoples are in neutral. Overall opinions of peoples on Swacha Bharat scheme is Positives.
Table VII Comparative Analysis of Kashmir twitter dataset
Dataset % Pos % Neg % Neu
Kashmir 2016 23.17 35 41.97
Kashmir 2017 20.74 31.58 47.8
Kashmir 2018 25.26 32.82 41.92
Figure 8. Comparative Analysis of Kashmir twitter Dataset
in above Comparative Analysis of Kashmir twitter dataset table in year 2016 23.17 % of tweets are categorized as positive, in year 2017 20.74 % of tweets are positive and in year 2018 25.26% of tweets are categorized as positive tweets in 2016 35
% of tweets categorized as negative and in year 2017 31.58 % of tweets are Negative, in year 2018 32.82 % of tweets are categorized negative and in year 2016 41.97 % of tweets are categorized as neutral and in year 2017 47.8 % of tweets are categorized as neutral and year 2018 41.92 % of tweets are categorized as neutral means in the year 2016 and year 2017 less peoples are positive for Kashmir and in year 2016 and year 2018 more peoples are negative in 2016 year peoples are having less Neutral opinion i.e. 41.97 % in year 2017 74.8%
peoples are more neutral opinions. Overall opinions of peoples on Kashmir scheme is Negatives.
Table VIII Comparative Analysis of Yoga Day twitter dataset
Figure 9. Comparative Analysis of Yoga Day twitter dataset
in above Comparative Analysis of Yoga Day twitter dataset table in year 2016 38.46 % of tweets are categorized as positive, in year 2017 41.54 % of tweets are positive, in year 2018 52.32 % of tweets are categorized as Positive, and in 2016 16.4 % of tweets categorized as negative and in year 2017 8.41 % of tweets are Negative, in year 2018 1.38 % of tweets are categorized as negative and in year 2016 45.14 % of tweets are categorized as neutral, in year 2017 50.03 % of tweets are categorized as neutral and in year 2018 46.3 tweets are categorized as neutral means in the year 2016 and year 2018 more peoples are positive for Yoga Day and in year 2017 and year 2018 less peoples are negative in 2016 year peoples are having less Neutral opinion i.e. 45.14 % in year 2017 50.03 % peoples are more neutral opinions. Overall opinions of peoples on Yoga Day scheme is Positives.
V. CONCLUSION
In this research paper, show new approach of prediction of 2019 Indian Election. In 2019 main elections agendas are probably GST, Demonetization, Digital India, Make in India, Startup India, Swacha Bharat, Kashmir and Yoga day. In this paper, we are worked on that schemes and predict the popularity of schemes using sentiment analysis score method.
Tweets are used to find out which scheme is most supported by the public. The paper has implemented and shown the results of the prediction of 2019 Indian Election. It can be seen that doing sentiment analysis on Twitter data for opinion mining can be used for prediction of election results. Here we are predicting new government in 2019 Election our observations are Demonetization and Kashmir are have negative opinions of peoples and GST, Digital India, Make in India, Startup India, Swacha Bharat and Yoga day have positive opinions. On the basis of this observations and evidences of dataset in 2019 same government is to be elected.
ACKNOWLEDGMENT
We are author’s sincerely thanks to Department of Computer Science and Information Technology, Dr. Babasaheb
Ambedkar Marathwada University for providing research laboratory facility for completion of this work. Also we would like to thanks Ministry of Electronics & Information Technology (MeitY), Government of India for provided Fellowship under VISVESVARAYA PHD SCHEME FOR ELECTRONICS & IT.
REFERENCES
[1] A.M. Abirami, V. Gayathri “A Survey on Sentiment Analysis Methods and Approach” 2016 Eighth International Conference on Advanced Computing (ICoAC) 19-21 Jan. 2017 Electronic ISBN:
978-1-5090-5888-4 Print on Demand (PoD) ISBN: 978-1-5090-5889- 1 DOI: 10.1109/ICoAC.2017.7951748 Page no-72 to 76.
[2] Santanu Mandal, Sumit Gupta “A Novel Dictionary-Based Classification Algorithm for Opinion Mining” Second International Conference on Research in Computational Intelligence and communication networks 23-25 Sept. 2016 Kolkata, India Electronic ISBN: 978-1-5090-1047-9CD-ROM ISBN: 978-1-5090-1045-5USB ISBN: 978-1-5090-1046-2 Print on Demand (PoD) ISBN: 978-1- 5090-1048-6DOI: 10.1109/ICRCICN.2016.7813652Page no- 175 to 180.
[3] P. Waila, Marisha, V.K. Singh and M.K. Singh “Evaluating Machine Learning and Unsupervised Semantic Orientation Approaches for Sentiment Analysis of Textual Reviews” 2012 IEEE International Conference on Computational Intelligence and Computing Research 18-20 Dec. 2012 Electronic ISBN: 978-1-4673- 1344-5 Print ISBN: 978-1-4673-1342-1 CD-ROM ISBN: 978-1- 4673-1343-8 Print on Demand (PoD) ISBN: 978-1-4673-1342-1 DOI: 10.1109/ICCIC.2012.6510235.Page no-01 to 06.
[4] Durgesh M. Sharma, Prof. Moiz M. Baig” Sentiment Analysis on Social Networking: A Literature Review” International Journal on Recent and Innovation Trends in Computing and Communication February 2015 Volume: 3 Issue: 2 ISSN: 2321-8169 page no-22-27.
[5] R. Nivedha and N. Sairam”A Machine Learning based Classification for Social Media Messages” Indian Journal of Science and Technology, Vol 8(16), DOI: 10.17485/ijst/2015/v8i16/63640, July 2015 ISSN (Print): 0974-6846 ISSN (Online) : 0974-5645 Page No-1 to 4.
[6] Harsha Sinha, Arashdeep Kaur “A Detailed Survey and Comparative Study of Sentiment Analysis Algorithms” IEEE 2nd International Conference on Communication, Control and Intelligent Systems (CCIS) 18-20 Nov. 2016Mathura, India Electronic ISBN:
978-1-5090-3210-5DOI: 10.1109/CCIntelS.2016.7878208Page no 94 to 98.
[7] Nien-Lin Hsueh”A Framework for Opinion Mining System with Design Pattern” International Computer Symposium 15-17 Dec.
2016Chiayi, Taiwan Electronic ISBN: 978-1-5090-3438-3DOI:
10.1109/ICS.2016.0128 Page no-626 to 631.
[8] Nipuna Upeka Pannala et.al “Supervised Learning Based Approach to Aspect Based Sentiment Analysis” IEEE International Conference on Computer and Information Technology (CIT) 8-10 Dec. 2016Nadi, Fiji Electronic ISBN: 978-1-5090-4314-9DOI 10.1109/CIT.2016.107 Page No-662-666.
[9] L B Shyamasundar, P Jhansi Rani” Twitter sentiment analysis with different feature extractors and dimensionality reduction using supervised learning algorithms” IEEE Annual India Conference (INDICON) 16-18 Dec. 2016 Bangalore, India Electronic ISBN: 978- 1-5090-3646-2 DOI: 10.1109/INDICON.2016.7839075 Page No-1-6.
[10] Aliza Sarlan1, Chayanit Nadam2, Shuib Basri3 “Twitter Sentiment Analysis” International Conference on Information Technology and Multimedia (ICIMU), November 18 – 20, 2014, Putrajaya, Malaysia ISBN, 978-1-4799-5423-0/14 Page No-212-216.
Dataset % Pos % Neg % Neu
Yoga Day 2016 38.46 16.4 45.14
Yoga Day 2017 41.54 8.41 50.03
Yoga Day 2018 52.32 1.38 46.3