IMPLEMENTASI ALGORITMA SVM DENGAN AUGMENTASI TEKS UNTUK ANALISIS SENTIMEN BERBASIS ASPEK PADA ULASAN GAME SONS OF THE FOREST

TITO ADHITYA PRATAMA, . (2026) IMPLEMENTASI ALGORITMA SVM DENGAN AUGMENTASI TEKS UNTUK ANALISIS SENTIMEN BERBASIS ASPEK PADA ULASAN GAME SONS OF THE FOREST. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

	Text COVER.pdf Download (891kB)
	Text BAB 1.pdf Download (425kB)
	Text BAB 2.pdf Restricted to Registered users only Download (933kB) | Request a copy
	Text BAB 3.pdf Restricted to Registered users only Download (988kB) | Request a copy
	Text BAB 4.pdf Restricted to Registered users only Download (2MB) | Request a copy
	Text BAB 5.pdf Restricted to Registered users only Download (389kB) | Request a copy
	Text Daftar Pustaka.pdf Download (381kB)
	Text LAMPIRAN.pdf Restricted to Registered users only Download (2MB) | Request a copy

Abstract

Penelitian ini bertujuan melakukan Aspect-Based Sentiment Analysis (ABSA) pada ulasan permainan video Sons of the Forest. Pemodelan sentimen membandingkan arsitektur Support Vector Machine (SVM) dengan berbagai fungsi kernel terhadap algoritma K-Nearest Neighbor (KNN) sebagai baseline. Klasifikasi dilakukan ke dalam kelas positif, netral, dan negatif pada lima aspek: Gameplay, Graphics, Performance, Story, dan Atmosphere. Ekstraksi fitur menggunakan Term Frequency-Inverse Document Frequency (TF-IDF), dipadukan dengan teknik Easy Data Augmentation (EDA) berbasis WordNet untuk mengatasi ketidakseimbangan kelas. Hasil pengujian menunjukkan SVM secara konsisten mengungguli KNN. Konfigurasi SVM kernel Linear menghasilkan performa klasifikasi paling stabil dengan perolehan F1-Score 92,6% pada aspek Gameplay. Augmentasi teks terbukti efektif meningkatkan kemampuan deteksi kelas minoritas pada SVM, namun menurunkan performa KNN secara drastis akibat meluasnya dimensi ruang fitur yang mengganggu perhitungan jarak antar dokumen. Lebih lanjut, pemetaan sentimen mengungkapkan apresiasi tinggi terhadap kualitas visual melalui 6.233 ulasan positif pada aspek Graphics, namun didominasi ketidakpuasan pada aspek Performance dengan 11.806 ulasan negatif terkait kendala optimasi teknis. ***** This study conducts Aspect-Based Sentiment Analysis (ABSA) on user reviews of the game Sons of the Forest. The modeling compares the Support Vector Machine (SVM) architecture across various kernel functions with the K-Nearest Neighbor (KNN) baseline algorithm. Sentiments are classified into positive, neutral, and negative classes across five aspects: Gameplay, Graphics, Performance, Story, and Atmosphere. The research employs Term Frequency-Inverse Document Frequency (TF-IDF) for feature extraction and Easy Data Augmentation (EDA) via WordNet to address class imbalance. The results demonstrate that SVM consistently outperforms KNN. The SVM Linear kernel yields the most stable classification performance, achieving a 92.6% F1-Score in the Gameplay aspect. Text augmentation is proven effective in improving minority class detection in SVM, but it drastically degrades KNN's performance due to increased feature dimensionality that disrupts document distance calculations. Furthermore, aspect-level mapping reveals high appreciation for visual quality with 6,233 positive reviews in the Graphics aspect, while exposing dominant dissatisfaction in the Performance aspect, accumulating 11,806 negative reviews regarding technical optimization issues.

Item Type:	Thesis (Sarjana)
Additional Information:	1). Dr. Widodo, S.Kom., M.Kom. ; 2). Murien Nugraheni, S.T., M.Cs.
Subjects:	Teknologi dan Ilmu Terapan > Teknik Komputer
Divisions:	FT > S1 Pendidikan Teknik Informatika Komputer
Depositing User:	Users 35421 not found.
Date Deposited:	12 Aug 2026 07:27
Last Modified:	13 Aug 2026 02:37
URI:	http://repository.unj.ac.id/id/eprint/69908

Actions (login required)

View Item