Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal Sistemasi: Jurnal Sistem Informasi
Saputra, Fendi Pradana
Unknown Affiliation
Computer Science & IT Electrical & Electronics Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
Combining SVM and Naive Bayes Models using a Soft Voting Approach for Sentiment Analysis of Tong Tji Tea House Saputra, Fendi Pradana; Suria, Ozzi
Sistemasi: Jurnal Sistem Informasi Vol 14, No 5 (2025): Sistemasi: Jurnal Sistem Informasi
Publisher : Program Studi Sistem Informasi Fakultas Teknik dan Ilmu Komputer

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.32520/stmsi.v14i5.5481

Abstract

In today’s digital technology and social media era, people are increasingly influenced to actively share reviews of restaurant services, expressing a wide range of customer opinions and perceptions. This study aims to analyze sentiment in Indonesian-language review texts using three machine learning models: Support Vector Machine (SVM), Naive Bayes (NB), and a combination of both through an Ensemble Soft Voting Classifier approach. The research focuses on user reviews of the Tong Tji Tea House, collected from the Google Maps platform, with sentiment data distributed as follows: positive (2,676 entries), neutral (670 entries), and negative (251 entries). The class imbalance poses a significant challenge in developing an optimal classification model. To address this, parameter optimization was carried out using the Grid Search method. The SVM model with a linear kernel and C=10 parameter achieved an accuracy of 0.9289 and a positive F1-score of 0.9289. The NB model recorded an accuracy of 0.8340 with an F1-score of 0.9102. Meanwhile, the Ensemble model with a soft voting approach and a 4:1 weight ratio (SVM:NB) demonstrated the best performance, achieving an accuracy of 0.9344 and a positive F1-score of 0.9750. These results indicate that the Ensemble method effectively enhances model accuracy and robustness in handling imbalanced data.
Co-Authors Suria, Ozzi