Analisis Sentimen Opini Masyarakat Terhadap Pembelajaran Daring Era Pandemi Dari Twitter Menggunakan Algoritma K-Nearest Neighbor (K-Nn)

INDRIANI, Theresia Ekawati (2023) Analisis Sentimen Opini Masyarakat Terhadap Pembelajaran Daring Era Pandemi Dari Twitter Menggunakan Algoritma K-Nearest Neighbor (K-Nn). Undergraduate thesis, Universitas Katolik Widya Mandira Kupang.

	Text ABSTRAK.pdf Download (335kB)
	Text BAB I.pdf Download (399kB)
	Text BAB II.pdf Restricted to Repository staff only Download (459kB)
	Text BAB III.pdf Restricted to Repository staff only Download (226kB)
	Text BAB IV.pdf Restricted to Repository staff only Download (649kB)
	Text BAB V.pdf Download (311kB)

Abstract

The spread of the Covid-19 virus in Indonesia in 2019 resulted in 463,000 people being confirmed positive and the death toll reaching 15,148 people. Online learning is a trending topic on Twitter because many people give their opinions regarding online learning. A collection of opinions on Twitter can be analyzed to find out the negative, positive and neutral sentiments contained in public opinion. This study uses the K-Nearest Neighbor (K-NN) method to analyze sentiment. The research data is a collection of 50,000 Indonesian language tweet data and after preprocessing the data becomes 9,000 data taken on November 17 2022 - December 14 2022 by utilizing the Twitter API and using the keywords "Online Learning" and "Online". Sentiment analysis results from applying the classification to public opinion data on online learning on Twitter social media that positive sentiments are superior to 4055 compared to 3264 negative sentiment classes and 1681 neutral sentiments. By looking at the results of the confusion matrix, the values for accuracy, precision, recall, and f-1 scores. Based on the value of K = 95 and the number of folds 20, the accuracy value is 95.5%, precision is 92%, recall is 92% and f1-score is 91.9%.

Actions (login required)

View Item