EVOLUSI : Jurnal Sains dan Manajemen
Vol 12, No 2 (2024): Jurnal Evolusi 2024

Analisa Sentimen Persepsi Masyarakat Terhadap Aplikasi Bea Cukai Mobile Menggunakan Algoritma Naive Bayes Dan K-Nearest Neighbors

Rousyati, Rousyati (Unknown)
Pratmanto, Dany (Unknown)
Widodo, Andrian Eko (Unknown)
Fatmawati, Kulum (Unknown)
Saputra, Rangga Diva (Unknown)



Article Info

Publish Date
01 Sep 2024

Abstract

AbstractThis study analyzes user sentiment towards the Bea Cukai Mobile application using Naive Bayes and K-Nearest Neighbors (KNN) algorithms. Data was collected from 600 reviews on Google Play Store, equally divided between positive and negative sentiments. After preprocessing, the data was analyzed using both algorithms. Results show that Naive Bayes outperformed with 79.96% accuracy, 87.76% recall, and 96.40% AUC, compared to KNN's 78.34% accuracy, 75.32% recall, and 92.20% AUC. However, KNN excelled in precision with 81.06% versus Naive Bayes' 77.21%. The study concludes that Naive Bayes is more effective in providing accurate classification and distinguishing between positive and negative classes, while KNN is more precise in predicting positive classes. These findings offer valuable insights into user perceptions of the Bea Cukai Mobile application and the effectiveness of algorithms in sentiment analysis.Keywords: Sentiment analysis, Bea Cukai Mobile application, Naive Bayes, K-Nearest Neighbors, Google Play Store, User reviewsAbstrakPenelitian ini menganalisis sentimen pengguna terhadap aplikasi Bea Cukai Mobile menggunakan algoritma Naive Bayes dan K-Nearest Neighbors (KNN). Data diperoleh dari 600 ulasan di Google Play Store, terbagi sama rata antara sentimen positif dan negatif. Setelah melalui tahap preprocessing, data dianalisis menggunakan kedua algoritma tersebut. Hasil menunjukkan bahwa Naive Bayes memiliki performa lebih baik dengan akurasi 79,96%, recall 87,76%, dan nilai AUC 96,40%, dibandingkan KNN dengan akurasi 78,34%, recall 75,32%, dan AUC 92,20%. Namun, KNN unggul dalam hal presisi dengan 81,06% dibanding Naive Bayes 77,21%. Penelitian ini menyimpulkan bahwa Naive Bayes lebih efektif dalam memberikan klasifikasi akurat dan membedakan kelas positif dan negatif, sementara KNN lebih tepat dalam memprediksi kelas positif. Hasil ini memberikan wawasan berharga tentang persepsi pengguna terhadap aplikasi Bea Cukai Mobile dan efektivitas algoritma dalam analisis sentimen.Kata kunci: Analisis sentimen, Aplikasi Bea Cukai Mobile, Naive Bayes, K-Nearest Neighbors, Google Play Store, Ulasan pengguna

Copyrights © 2024






Journal Info

Abbrev

evolusi

Publisher

Subject

Computer Science & IT

Description

The aim of this journal is to publish high-quality articles dedicated to all aspects of the latest outstanding developments in the fields of science, management and informatics in all settings. The scope of this journal encompasses the applications of the development of science, computer science, ...