An LDA Analysis for Topic Modeling of the RESTI Journal

Abstract

This research applies latent Dirichlet allocation (LDA) modeling to uncover and analyze key topics and trends in the RESTI journal (Rekayasa Sistem dan Teknologi Informasi) journal from 2018-2022. A corpus of 594 articles was compiled and preprocessed to extract keywords/phrases. An optimized 10-topic LDA model clustered articles based on abstract texts. Quantitative and qualitative techniques analyzed the resulting topics, proportions, dynamics, and relationships. Image classification, social modeling, student systems, service applications, and digital studies emerged as dominant themes. Topic modeling revealed dramatic uptrends in AI and computer vision, mirroring explosive progress in deep learning. Disciplinary trends quantified shifts toward core computer science along with declines in social science and education. Keyword frequencies traced the rapid emergence of AI terminology. The demonstrated framework that combines metadata analysis, probabilistic topic modeling, temporal topic dynamics, and keyword trends provides a robust methodology to longitudinally analyze knowledge landscapes from research corpora. Computational analytical approaches fill a gap by showcasing versatile techniques for uncovering latent semantic patterns and research trajectories from scientific publications, offering novel tools for mapping the evolution of research themes over time. Beyond the specific context of RESTI, the presented text mining methodology has broad applications to reveal insights from diverse cross-disciplinary publication archives.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 1 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.