Riyanni, Nasyawa Adesty and , Khanun Roisatul Ummah, S.Tr.T., M.Tr.Kom. (2026) Studi Komparatif Metode TF-IDF, Word2Vec Dan Hibrida Menggunakan Klasifikasi Support Vector Machine, Naïve Bayes Dan Long Short-Term Memory Pada Data ACAB (All Cops Are Bastards). Skripsi thesis, Universitas Muhammadiyah Surakarta.
![[img]](https://eprints.ums.ac.id/style/images/fileicons/application_pdf.png) |
PDF (Naskah Publikasi)
NASKAH PUBLIKASI.pdf Download (585kB) |
|
PDF (Surat Pernyataan Publikasi)
Surat Pernyataan.pdf Restricted to Repository staff only Download (599kB) |
Abstract
The ACAB (All Cops Are Bastards) issue has again become a hot topic of discussion after the tragedy during the demonstration on August 28, 2025, which claimed the life of an online motorcycle taxi driver, Affan Kurniawan, who was run over by a Brimob car. This incident was widely discussed on various social media platforms, especially on X, many people criticized the actions of the police officers who were considered to reflect abuse of power. This study aims to analyze comments on X tweets related to the ACAB (All Cops Are Bastards) issue using the word ACAB on August 15, 2025, to September 15, 2025. The dataset used in the study consists of 7,634 tweets which will then be processed using TF-IDF, Word2Vec and hybrid Word2Vec-TF-IDF text representations and classified using the Support Vector Machine (SVM), Naïve Bayes (NB) and Long Short-Term Memory (LSTM) methods. The analysis results show 74.8% negative sentiment, 19.4% neutral sentiment, and 5.8% positive sentiment. These results indicate a significant class imbalance between positive and negative. This data distribution arises because the ACAB slogan has connotations of criticism and resistance to the police institution. The evaluation results show that the hybrid representation has more stable and consistent performance results than a single representation. The best combination of representation and classification results is the Word2Vec-TF-IDF hybrid representation with SVM classification where the accuracy is 0.76, precision 0.72, recall 0.76, and F1-Score 0.72 is better than other combinations. This test proves that the Word2Vec-TF-IDF hybrid text representation has a better opportunity to be developed in sentiment analysis.
| Item Type: | Thesis (Skripsi) |
|---|---|
| Uncontrolled Keywords: | Hybrid Word2Vec-TF-IDF, LSTM, Naïve Bayes, LSTM, SVM, TF-IDF, Word2Vec |
| Subjects: | H Social Sciences > HE Communications > Social Media > Twitter (X) T Technology > T Technology (General) T Technology > Technical Information |
| Divisions: | Fakultas Komunikasi dan Informatika > S1 Teknik Informatika |
| Depositing User: | NASYAWA ADESTY RIYANI |
| Date Deposited: | 24 Feb 2026 03:42 |
| Last Modified: | 24 Feb 2026 03:42 |
| URI: | http://eprints.ums.ac.id/id/eprint/143369 |
Actions (login required)
 |
View Item |
