Sentiments Analysis for Governor of East Java 2018 in Twitter

Buntoro, Ghulam Asrofi (2019) Sentiments Analysis for Governor of East Java 2018 in Twitter. Journal Publication & Informatics Engineering Research, 3 (2). pp. 49-55. ISSN 2541-044X

	Text Sentiments Analysis for Governor of East Java 2018 in Twitter.pdf Download (187kB)
	Text Sentiments Analysis for Governor of East Java 2018 in Twitter.pdf Download (2MB)
	Text Sentiments Analysis for Governor of East Java 2018 in Twitter.pdf Download (2MB)

Abstract

East Java Governor Election 2018 is also felt in the virtual world especially Twitter social media. All people freely argue about their respective governor candidates, the memorandum raises many opinions, not only positive or neutral but also negative opinions. Media growth is so rapid, revealing a lot of online media from the news media to social media. Social media alone is Facebook, Twitter, Path, Instagram, Google+, Tumblr, Linkedin and many more. Today's social media is not only used as a means of friendship or making friends, but also for other activities. Promos of trading or buying and selling, until political party promos or campaigns of candidates for regents, governors, legislative candidates until presidential candidates. The research objective is to conduct a Method with Data Prepocessing Selection comments, Cleansing, Parsing, Normalizing, Tokeninzing, Load Dictionary, Weighting and Classification of analyzing the sentiments of 2018 East Java Governor candidates on Twitter social media with optimal and maximum optimization. While the benefits are to help the community conduct research on opinions on twitter which contains positive, neutral or negative sentiments. Analysis of the sentiments of East Java Governor candidates in 2018 on twitter social media using non-conventional processes that save costs, time and effort. The results of Khofifah's dataset are 77% accuracy, 79.2% precision value, 77% recall value, 98.6% TP rate and 22.2% TN rate. For the results of Gus dataset, the accuracy is 76%, the precision value is 74.4%, the recall value is 76%, the TP rate is 93.8% and the TN rate is 52.9%.

Item Type:	Article
Uncontrolled Keywords:	sentiment analysis, governor, lexicon based features, naïve bayes classifier
Subjects:	T Technology > T Technology (General)
Depositing User:	Library Umpo
Date Deposited:	21 Oct 2019 07:45
Last Modified:	12 Dec 2019 02:48
URI:	http://eprints.umpo.ac.id/id/eprint/5075

Actions (login required)

View Item