%0 Thesis
%9 Skripsi
%A Fadhilah, Anas
%A Universitas Pakuan,
%A Fakultas Matematika dan Ilmu Pnegetahuan Alam,
%A Program Studi Ilmu Komputer,
%B Fakultas Matematika dan Pengetahuan Alam
%D 2025
%F eprintsunpak:11071
%I Universitas Pakuan
%T Analisis Sentimen Pada Hashtag #Kabupaten Aja dulu Menggunakan Metode Support Vector Machine (SVM) Dengan Word2vec Bidirectional Encoder Representations From Transformers (BERT)
%U http://eprints.unpak.ac.id/11071/
%X Social media has become a major platform for expressing public opinion regarding social, political, and  economic issues. One of the viral phenomena currently discussed in Indonesia is the hashtag #KaburAjaDulu,  which reflects public concerns regarding employment opportunities, economic instability, and social  conditions. This study aims to compare the performance of Support Vector Machine (SVM) combined with  Word2Vec and Bidirectional Encoder Representations from Transformers (BERT) in sentiment analysis of  tweets containing the #KaburAjaDulu hashtag on X (Twitter). The dataset was collected using web scraping  techniques and processed through several preprocessing stages, including cleaning text, case folding, slang  normalization, tokenization, stopword removal, and sentence reconstruction. The research adopted the Cross  Industry Standard Process for Data Mining (CRISP-DM) framework consisting of business understanding,  data understanding, data preparation, modeling, evaluation, and deployment stages. In the first approach,  Word2Vec with the Skip-gram architecture was used to generate word embeddings which were classified  using SVM with a Radial Basis Function (RBF) kernel. The second approach employed IndoBERT for  contextual sentiment classification. Model evaluation was conducted using accuracy, precision, recall, F1-  score, and confusion matrix metrics. Experimental results demonstrated that BERT achieved superior  classification performance compared with SVM combined with Word2Vec due to its contextual  understanding capability. The findings indicate that transformer-based models provide more accurate  sentiment classification for Indonesian social media text, especially in detecting contextual and implicit  sentiment expressions.  Keywords: Sentiment Analysis; Support Vector Machine; Word2Vec; BERT; Twitter; Natural Language  Processing; CRISP-DM; #KaburAjaDulu