IMPLEMENTASI METODE ENSEMBLE MAJORITY VOTE PADA ALGORITMA NAIVE BAYES DAN RANDOM FOREST UNTUK ANALISIS SENTIMEN TWITTER HARGA TIKET PESAWAT DOMESTIK

Triyana, Dimas (2024) IMPLEMENTASI METODE ENSEMBLE MAJORITY VOTE PADA ALGORITMA NAIVE BAYES DAN RANDOM FOREST UNTUK ANALISIS SENTIMEN TWITTER HARGA TIKET PESAWAT DOMESTIK. Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 20081010049.-cover.pdf Download (3MB)
	Text (Bab 1) 20081010049.-bab 1.pdf Download (950kB)
	Text (Bab 2) 20081010049.-bab 2.pdf Restricted to Repository staff only until 19 July 2026. Download (2MB)
	Text (Bab 3) 20081010049.-bab 3.pdf Restricted to Repository staff only until 19 July 2026. Download (9MB)
	Text (Bab 4) 20081010049.-bab 4.pdf Restricted to Repository staff only until 19 July 2026. Download (10MB)
	Text (Bab 5) 20081010049.-bab 5.pdf Download (318kB)
	Text (Daftar Pustaka) 20081010049.-daftar pustaka.pdf Download (676kB)

Abstract

Transportation is a crucial component in supporting the economic growth of a country. The domestic aviation industry in Indonesia, in particular, has experienced rapid growth after being impacted by the Covid-19 pandemic. However, the increase in passenger numbers has led to differing opinions regarding the rising prices of domestic airline tickets, which has become a critical factor in consumer purchasing decisions. Sentiment analysis through social media platforms like Twitter provides a solution to understanding public opinion on domestic airline ticket prices. This study focuses on using Naive Bayes and Random Forest, followed by the implementation of Ensemble Majority Voting for sentiment analysis on tweets related to domestic flight ticket prices. In the first scenario, Naive Bayes was applied using a training and testing data ratio of 80:20, achieving an accuracy of 70.31% after applying SMOTE to address data labeling imbalance. In the second scenario, with a 70:30 ratio, an accuracy of 71.75% was obtained. The Random Forest method was also used as a comparison algorithm. In the first scenario, Random Forest achieved an accuracy of 84.91%, and in the second scenario, 85.71%. Combining the two algorithms using ensemble majority voting resulted in accuracies of 85.88% and 85.87% for the two scenarios, indicating an improvement in prediction accuracy compared to the individual models. Thus, the ensemble majority vote method is deemed effective in enhancing the accuracy of sentiment analysis regarding domestic airline ticket prices. Public opinion on Twitter concerning domestic airline ticket prices is predominantly negative.

Item Type:

Thesis (Undergraduate)

Contributors: