Identifikasi Topik Pada Opini Publik Tentang Coretax Administration System (CTAS) Di Sosial Media Menggunakan Metode Bertopic

Nugroho, Satria Raditya (2026) Identifikasi Topik Pada Opini Publik Tentang Coretax Administration System (CTAS) Di Sosial Media Menggunakan Metode Bertopic. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043221001-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Transformasi digital pada sektor perpajakan diwujudkan melalui penerapan Core Tax Administration System (CTAS) oleh Direktorat Jenderal Pajak untuk meningkatkan efisiensi dan kemudahan layanan perpajakan daring. Penelitian ini menganalisis persepsi publik terhadap CTAS berdasarkan opini pengguna di media sosial X dan YouTube menggunakan pendekatan analisis sentimen dan pemodelan topik. Data diperoleh melalui teknik crawling terhadap 5.000 komentar periode Januari–Juni 2025. Analisis sentimen dilakukan menggunakan model IndoBERT, sedangkan pemodelan topik menggunakan metode BERTopic dengan evaluasi berbasis coherence score. Hasil penelitian menunjukkan bahwa sentimen negatif secara signifikan mendominasi pada kedua platform media sosial. Pada platform X, sentimen negatif terutama berkaitan dengan permasalahan administratif dan operasional, seperti kesulitan pendaftaran dan login akun, pengelolaan NPWP, serta kendala dalam proses pelaporan dan pembayaran pajak. Sementara itu, pada platform YouTube, sentimen negatif lebih banyak dipicu oleh masalah teknis aplikasi, khususnya pada tahap aktivasi akun, verifikasi identitas, dan kegagalan akses sistem. Model IndoBERT menunjukkan performa klasifikasi yang sangat baik dengan nilai F1-score sebesar 0,9842 (negatif) dan 0,8807 (positif) pada YouTube, serta 0,9481 (negatif) dan 0,7034 (positif) pada platform X, yang mengindikasikan kemampuan model dalam menangkap konteks bahasa Indonesia secara akurat. Pemodelan topik menggunakan BERTopic mengungkap perbedaan karakter diskusi antarplatform, di mana platform X menghasilkan topik yang lebih terfokus dan koheren dengan nilai coherence score 0,605–0,629, sedangkan YouTube menghasilkan topik yang lebih beragam dan naratif dengan nilai 0,388–0,410. Temuan ini menunjukkan bahwa isu teknis dan administratif menjadi perhatian utama publik dalam implementasi CTAS dan dapat digunakan sebagai dasar evaluasi peningkatan kualitas layanan perpajakan digital.
===================================================================================================================================
Digital transformation in the taxation sector is implemented through the Core Tax Administration System (CTAS) by the Directorate General of Taxes to improve the efficiency and accessibility of online tax services. This study analyzes public perceptions of CTAS based on user opinions on X and YouTube using sentiment analysis and topic modeling approaches. The data were collected through web crawling of 5,000 comments from January to June 2025. Sentiment analysis was conducted using the IndoBERT model, while topic modeling employed the BERTopic method evaluated using coherence scores. The results indicate that negative sentiment dominates on both social media platforms. On X, negative sentiment is primarily associated with administrative and operational issues, such as difficulties in account registration and login, NPWP management, and obstacles in tax reporting and payment processes. Meanwhile, on YouTube, negative sentiment is mainly driven by technical application issues, particularly during account activation, identity verification, and system access failures. The IndoBERT model demonstrates strong classification performance, achieving F1-scores of 0.9842 (negative) and 0.8807 (positive) on YouTube, and 0.9481 (negative) and 0.7034 (positive) on X, indicating its effectiveness in capturing Indonesian language context. Topic modeling using BERTopic reveals differences in discussion characteristics across platforms, where X produces more focused and coherent topics with coherence scores ranging from 0.605 to 0.629, while YouTube generates more diverse and narrative topics with scores between 0.388 and 0.410. These findings indicate that technical and administrative issues are the primary public concerns in the implementation of CTAS and can serve as a basis for improving digital tax service quality.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Analisis Sentimen, BERTopic, Core Tax Administration System (CTAS), IndoBERT, Pemodelan Topik. Core Tax Administration System, sentiment analysis, topic modeling, IndoBERT, BERTopic, social media.
Subjects:	T Technology > T Technology (General) > T57.8 Nonlinear programming. Support vector machine. Wavelets. Hidden Markov models.
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Satria Raditya Nugroho
Date Deposited:	13 Feb 2026 06:09
Last Modified:	13 Feb 2026 06:09
URI:	http://repository.its.ac.id/id/eprint/132434

Actions (login required)

View Item