SOESANTO, IVAN CHRISTIAN (2024) SENTIMENT ANALYSIS ON INSTAGRAM COMMENTS USING TWO COMPARISON METHOD BASED ON LEXICON BASED AND NAIVE BAYES. S1 thesis, UNIVERSITAS KATOLIK SOEGIJAPRANATA.

Preview	Text 21.K1.0006_IVAN CHRISTIAN SOESANTO_COVER.pdf Download (437kB) | Preview
	Text 21.K1.0006_IVAN CHRISTIAN SOESANTO_BAB I.pdf Restricted to Registered users only Download (410kB)
	Text 21.K1.0006_IVAN CHRISTIAN SOESANTO_BAB II.pdf Restricted to Registered users only Download (481kB)
	Text 21.K1.0006_IVAN CHRISTIAN SOESANTO_BAB III.pdf Restricted to Registered users only Download (744kB)
	Text 21.K1.0006_IVAN CHRISTIAN SOESANTO_BAB IV.pdf Restricted to Registered users only Download (758kB)
	Text 21.K1.0006_IVAN CHRISTIAN SOESANTO_BAB V.pdf Restricted to Registered users only Download (409kB)
Preview	Text 21.K1.0006_IVAN CHRISTIAN SOESANTO_DAPUS.pdf Download (531kB) | Preview
	Text 21.K1.0006_IVAN CHRISTIAN SOESANTO_LAMP.pdf Restricted to Registered users only Download (571kB)

Abstract

Sentiment analysis is a very important tool for understanding public opinion and serves to share user experiences, especially in the context of social media. This thesis discusses the task of sentiment analysis using two comparisons of Lexicon Based and Naive Bayes methods to analyze comments from Instagram. The dataset used for this project was taken from Kaggle with keyword sentiment analysis on the Indonesian language Instagram application collected.In several studies that have been carried out, it turns out that the results that researchers can get or prove are that the Naive Bayes Algorithm gets a high Acccuracy value, namely 82% and compared to using a Lexicon Based Algorithm using the vader Dictionary, Where the Accuracy in Indonesian gets a result of 50% and Translated into English the Accuracy obtained was 60%. Therefore, for research in sentiment analysis on Instragram Comments, the Naive Bayes Algorithm is very suitable and is evaluated with Cross - Validation.

Actions (login required)

View Item