Article Per Year (5 Year)
p-Index From 2020 - 2025
0.659
P-Index
This Author published in this journals
All Journal Journal of Informatics Management and Information Technology
Yulianto Triwahyuadi Polly
Unknown Affiliation
Computer Science & IT

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

Found 1 Documents
Search
Journal : Journal of Informatics Management and Information Technology

 1 
Analisis Sentimen Terhadap Data Komentar Publik Mengenai Isu UU Pilkada 2024 Menggunakan Metode Naïve Bayes dan K-Nearest Neighbor Sebastianus Adi Santoso Mola; Yulianto Triwahyuadi Polly; Atok, Yosefa Carela
Journal of Informatics Management and Information Technology Vol. 5 No. 3 (2025): July 2025
Publisher : Forum Kerjasama Pendidikan Tinggi (FKPT)

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/jimat.v5i3.514

Abstract

The 2024 Regional Head Election Law (UU Pilkada) has become an important issue widely discussed in Indonesia, especially on the social media platform X. Various public comments related to this issue contain positive, negative, and neutral sentiments, reflecting public perceptions. This study aims to analyze the sentiment of public comments on the 2024 UU Pilkada using two machine learning methods: Naïve Bayes and K-Nearest Neighbor (K-NN). The dataset consists of 3864 comments divided into three sentiment classes: 1477 negative comments, 1385 neutral comments, and 1002 positive comments, all of which have undergone text preprocessing. Evaluation was conducted using k-fold cross-validation (k=10). The test results show that the Naïve Bayes method achieves the highest accuracy of 63.47%, while K-NN reaches 56.73%. The precision for negative sentiment is 56.84%, meaning that about 43% of the comments predicted as negative by the model are actually not negative. The recall for negative sentiment is 45.45%, indicating that the model only captures less than half of the actual negative comments. For neutral sentiment, the precision of 60.71% and recall of 66.23% suggest that the model performs fairly well in recognizing neutral comments, although there is still a 39.29% error. For positive sentiment, the precision of 55.55% and recall of 57.63% indicate errors in classifying positive comments. Overall, while the model can correctly classify a portion of the data, there is potential to improve accuracy for both the negative and positive classes.
Co-Authors Adriana Fanggidae Adriana Fanggidae Atok, Yosefa Carela Nelci Dessy Rumlaklak Rizal Pebianto Sebastianus Adi Santoso Mola