Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal Journal of applied statistics and data mining
Waldi Darmansyah
Unknown Affiliation
Computer Science & IT Decision Sciences, Operations Research & Management Economics, Econometrics & Finance Social Sciences

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

Found 1 Documents
Search

 1 
Comparison of Feature Extraction for Sentiment Analysis using Support Vector Machine Algorithm Waldi Darmansyah; Herman Yuliansyah 
Journal of Applied Statistics and Data Mining Vol. 6 No. 1 (2025): Journal Applied Statistics and Data Mining
Publisher : Institut Teknologi Statistika dan Bisnis Muhammadiyah Semarang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.63229/jasdm.v6i1.86

Abstract

Background: The Directorate General of Customs and Excise (DJBC), as the regulator of goods at airports, frequently receives passenger complaints about baggage inspections. A recent study (2023) showed a 25% increase in complaints via social media, but no research has compared feature extraction techniques for this specific sentiment analysis. Objective: This study aims to compare the performance of the BoW and TF-IDF methods in sentiment analysis of DJBC inspection complaints, develop an SVM model for sentiment classification, and identify passenger sentiment patterns from Twitter data. Methods: This quantitative research analyzed 4,215 tweets about DJBC from January to June 2023. The stages included: text preprocessing, feature extraction (BoW and TF-IDF), classification with SVM, and evaluation using accuracy, precision, recall, and F1-score. Results: The TF-IDF model achieved 91.3% accuracy (91% precision, 89% recall, and 90% F1-score), while the BoW model achieved 91.1% accuracy (92% precision, 90% recall, and 91% F1-score). The analysis showed that the BoW model was superior in capturing the nuances of complaints. Conclusion: Despite minimal accuracy differences, BoW was more effective for sentiment analysis of DJBC audit complaints. These findings recommend improving officer training on the most frequently complained-about aspects. Further research could test combinations with word embedding or transformers.
Co-Authors Herman YuliansyahÂ