ANALISIS SENTIMEN LEMBAGA SURVEI PADA PEMILIHAN PRESIDEN 2024 MENGGUNAKAN SUPPORT VECTOR MACHINE

ALISYA RAHMAWANTI, . (2026) ANALISIS SENTIMEN LEMBAGA SURVEI PADA PEMILIHAN PRESIDEN 2024 MENGGUNAKAN SUPPORT VECTOR MACHINE. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

	Text COVER.pdf Download (2MB)
	Text BAB I.pdf Download (419kB)
	Text BAB II.pdf Restricted to Registered users only Download (970kB) | Request a copy
	Text BAB III.pdf Restricted to Registered users only Download (437kB) | Request a copy
	Text BAB IV.pdf Restricted to Registered users only Download (1MB) | Request a copy
	Text BAB V.pdf Restricted to Registered users only Download (325kB) | Request a copy
	Text DAFTAR PUSTAKA.pdf Download (304kB)
	Text LAMPIRAN.pdf Restricted to Registered users only Download (545kB) | Request a copy

Abstract

Media sosial menjadi salah satu sarana utama masyarakat dalam menyampaikan opini terhadap berbagai isu politik, termasuk terkait lembaga survei pada pemilihan presiden tahun 2024. Opini yang berkembang di media sosial dapat mencerminkan persepsi publik yang beragam, sehingga diperlukan analisis sentimen untuk mengelompokkan opini tersebut secara sistematis. Penelitian ini bertujuan untuk mengklasifikasikan sentimen masyarakat terhadap lembaga survei Lembaga Survei Indonesia (LSI), Center for Strategic and International Studies (CSIS), dan Poltracking Indonesia berdasarkan data tweet dari media sosial X. Metode yang digunakan dalam penelitian ini adalah text mining dengan tahapan text preprocessing, pelabelan sentimen menggunakan pendekatan lexicon-based, serta ekstraksi fitur menggunakan Term Frequency–Inverse Document Frequency (TF-IDF). Proses klasifikasi sentimen dilakukan menggunakan metode Support Vector Machine (SVM) multi-class dengan pendekatan One Against All. Untuk memperoleh parameter terbaik, digunakan GridSearchCV. Hasil penelitian menunjukkan bahwa sentimen positif mendominasi pembahasan terhadap ketiga lembaga survei, diikuti oleh sentimen negatif dan netral. SVM dengan kernel Radial Basis Function (RBF) menghasilkan performa yang lebih baik dibandingkan kernel linear berdasarkan nilai akurasi, presisi, recall, dan F-score. Dengan demikian, metode SVM dengan kernel RBF dinilai efektif dalam mengklasifikasikan sentimen masyarakat terhadap lembaga survei di media sosial X. ***** Social media has become one of the main platforms for the public to express opinions on various political issues, including survey institutions in 2024 presidential elections. Opinions expressed on social media reflect diverse public perceptions, making sentiment analysis necessary to systematically classify such opinions. This study aims to classify public sentiment toward survey institutions, namely Lembaga Survei Indonesia (LSI), Center for Strategic and International Studies (CSIS), and Poltracking Indonesia, based on tweet data from social media X. The method used in this study is text mining, which includes text preprocessing, sentiment labeling using a lexicon-based approach, and feature extraction using Term Frequency–Inverse Document Frequency (TF-IDF). Sentiment classification is performed using a multi-class Support Vector Machine (SVM) with a One Against All approach. To obtain the optimal parameters, GridSearchCV is applied. The results show that positive sentiment dominates discussions related to the three survey institutions, followed by negative and neutral sentiments. Based on model performance evaluation, SVM with a Radial Basis Function (RBF) kernel achieves better performance than the linear kernel in terms of accuracy, precision, recall, and F-score. Therefore, the SVM method with an RBF kernel is considered effective for classifying public sentiment toward survey institutions on social media X.

Item Type:	Thesis (Sarjana)
Additional Information:	1). Prof. Dr. Ir. Bagus Sumargo, M.Si. ; 2). Qorry Meidianingsih, M.Si.
Subjects:	Sains > Statistika
Divisions:	FMIPA > S1 Statistika
Depositing User:	Alisya Rahmawanti .
Date Deposited:	20 Aug 2026 06:32
Last Modified:	20 Aug 2026 06:32
URI:	http://repository.unj.ac.id/id/eprint/71964

Actions (login required)

View Item