cover
Contact Name
Verdi Yasin
Contact Email
verdiyasin29@gmail.com
Phone
+62213905050
Journal Mail Official
jmijayakarta@jayakarta.ac.id
Editorial Address
Jalan Salemba I No.10 Kel.Kenari, Kec.Senen, Jakarta Pusat 10430 Indonesia
Location
Kota adm. jakarta pusat,
Dki jakarta
INDONESIA
Jurnal Manajemen Informatika Jayakarta
ISSN : 27465985     EISSN : 27970930     DOI : 10.52362
Terbitan berkala ini bertujuan untuk menerbit hasil pemikiran ilmiah dan hasil penelitian yang dapat dipertanggung jawabkan oleh seorang peneliti/Author. Terbitan berkala ini juga konsentrasi dalam bidang Ilmu Komputer, Teknologi Informasi, Sistem Informasi, Rekayasa Perangkat Lunak dan Data Sains.
Articles 181 Documents
EVALUASI SENTIMEN MASYARAKAT TERHADAP KEBIJAKAN SUBSIDI KENDARAAN LISTRIK DI INDONESIA DENGAN PENDEKATAN INSET LEXICON, WORD EMBEDDING, DAN ALGORITMA SUPPORT VECTOR MACHINE Ridwan Ridwan; Hendarman Lubis
Jurnal Manajamen Informatika Jayakarta Vol 6 No 2 (2026): JMI Jayakarta (April 2026)
Publisher : Sekolah Tinggi Manajemen Informatika dan Komputer Jayakarta

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.52362/jmijayakarta.v6i2.2364

Abstract

The electric vehicle subsidy policy in Indonesia is one of the government's efforts to promote environmentally friendly energy usage and reduce carbon emissions. However, the implementation of this policy has generated diverse public responses, which can be analyzed through social media platforms. This study aims to evaluate public sentiment toward the electric vehicle subsidy policy in Indonesia using the InSet Lexicon approach, Word Embedding (Word2Vec), and the Support Vector Machine (SVM) algorithm. The dataset was collected from Twitter through a crawling process based on relevant keywords, resulting in 1,000 tweets. The research stages include text preprocessing, sentiment labeling using InSet Lexicon, feature extraction using Word2Vec, and classification using SVM. The results show that sentiment distribution consists of 45% positive, 35% negative, and 20% neutral. The classification model achieved an accuracy of 86%, precision of 83%, recall of 81%, and an F1-score of 82%. These results indicate that the proposed approach is effective in classifying sentiment. Furthermore, the use of Word Embedding improves text representation quality, which contributes to better model performance. This study provides insights into public perception and can serve as a reference for evaluating public policies.