Analisis Jejaring Peneliti Dan Topik Penelitian Pada Artikel Ilmiah Menggunakan Neo4J: Studi Kasus Bisnis Digital

Wijaya, Audrey Sasqhia (2026) Analisis Jejaring Peneliti Dan Topik Penelitian Pada Artikel Ilmiah Menggunakan Neo4J: Studi Kasus Bisnis Digital. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5025221055-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Transformasi digital dan keberlanjutan bisnis menghasilkan ribuan publikasi ilmiah setiap tahun di IEEE Xplore dan Scopus. Mengkaji banyak makalah secara manual membutuhkan berbulan-bulan, sehingga pendekatan Computational Systematic Literature Review (CSLR) berbasis pemodelan topik diperlukan. Tidak seperti LDA dan NMF yang berbasis frekuensi kata, BERTopic menangkap makna kontekstual menggunakan model bahasa transformer. Namun, integrasinya dengan analisis jaringan kolaborasi melalui knowledge graph pada domain digitalisasi bisnis dan transformasi digital masih terbatas. Penelitian ini menggunakan 8.782 artikel dari IEEE Xplore dan Scopus (2015–2025) melalui tahap pengujian perbandingan komposisi korpus, perbandingan tujuh model embedding, optimasi hyperparameter pada UMAP dan HDBSCAN, perbandingan LDA/NMF/BERTopic berdasarkan koherensi C_{v}, serta pembangunan knowledge graph Neo4j yang dianalisis melalui metrik sentralitas dan deteksi komunitas Louvain. BERTopic dengan embedding MiniLM menghasilkan koherensi tertinggi (C_{v}=0,8680, outlier 0,178), mengidentifikasi 16 topik utama dan 171 subtopik. Knowledge graph mencakup 33.977 simpul dan 131.223 relasi yang menghubungkan 7.204 makalah, 21.577 penulis, 4.876 institusi, dan 115 negara dengan 4.830 komunitas mengikuti batas institusional dan regional. Tema Keberlanjutan Bisnis mencatat pertumbuhan tertinggi (159,59%) dan Model Bisnis Berkelanjutan menggeser Transformasi Digital sebagai topik terbanyak diteliti pada 2025; subtopik tercepat tumbuh: CSR & Etika Bisnis (+380%) dan Tata Kelola AI (+240%).

===============================================================================================================================

Digital transformation and business sustainability generate thousands of scientific publications annually in IEEE Xplore and Scopus. Manually reviewing this volume of literature takes months, making Computational Systematic Literature Review (CSLR) via topic modeling essential. Unlike LDA and NMF which rely on word frequency, BERTopic captures contextual meaning using transformer-based language models. However, integrating this approach with collaboration network analysis through a knowledge graph in this domain remains limited. This study uses 8,782 articles from IEEE Xplore and Scopus (2015–2025) across five stages: corpus composition comparison, selection from seven embedding models, hyperparameter optimization UMAP and HDBSCAN, comparison of LDA/NMF/BERTopic using coherence C_{v}, and construction of a Neo4j knowledge graph analyzed through centrality metrics and Louvain community detection. BERTopic with MiniLM embedding achieved the highest coherence (C_{v}=0.868, outlier rate 0.178), identifying 16 parent topics and 171 subtopics. The knowledge graph comprises 33,977 nodes and 131,223 edges connecting 7,204 papers, 21,577 authors, 4,876 institutions, and 115 countries, with 4,830 communities following institutional and regional boundaries. Business Sustainability recorded the highest growth (159.59%) and Sustainable Business Models surpassed Digital Transformation as the most researched topic in 2025; fastest-growing subtopics: CSR & Business Ethics (+380%) and AI Governance (+240%).

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Pemodelan Topik, BERTopic, Knowledge Graph, Neo4j, Transformasi Digital, Keberlanjutan Bisnis, Computational Systematic Literature Review, Topic Modeling, Digital Transformation, Business Sustainability
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Audrey Sasqhia Wijaya
Date Deposited:	23 Jul 2026 14:36
Last Modified:	23 Jul 2026 14:36
URI:	http://repository.its.ac.id/id/eprint/136639

Actions (login required)

View Item