<![CDATA[The rapid growth of cybercrime in Indonesia has significantly transformed prosecutorial narratives in criminal justice proceedings. This study aims to examine the evolution of prosecutors' narratives in cybercrime cases using a Natural Language Processing (NLP)-based text mining approach. A qualitative-computational research design was employed by analyzing cybercrime court decisions published in the Supreme Court of the Republic of Indonesia Decision Directory from 2009 to 2025. The analytical framework consisted of document collection, text extraction, preprocessing, descriptive corpus analysis, Term Frequency–Inverse Document Frequency (TF-IDF) weighting, Latent Dirichlet Allocation (LDA) topic modeling, temporal topic evolution analysis, and co-occurrence network analysis to identify relationships among legal concepts. The findings reveal a substantial transformation in prosecutorial narratives from conventional evidence-oriented arguments toward more sophisticated reasoning emphasizing electronic evidence, digital forensics, digital evidence governance, personal data protection, and inter-agency collaboration in cybercrime prosecution. Temporal analysis demonstrates the emergence of new legal themes corresponding to regulatory developments and the increasing complexity of cybercrime. Furthermore, the co-occurrence network identifies electronic evidence, cybersecurity, and criminal liability as the central concepts shaping prosecutorial arguments. This study contributes methodologically by integrating text mining techniques into legal research and provides empirical evidence to support the development of more adaptive prosecution policies in addressing cybercrime in Indonesia.]]>
