Klasifikasi Dokumen Publik Berbasis NLP: Otomatisasi Proses Informasi Menuju Keterbukaan Data yang Adaptif dan Transparan

Retnowati Retnowati (Universitas Stikubank)
Veronica Lusiana (Universitas Stikubank)
Eko Nur Wahyudi (Universitas Stikubank)

Jurnal Teknik Informatika dan Teknologi Informasi
Vol. 5 No. 2 (2025): Agustus: Jurnal Teknik Informatika dan Teknologi Informasi

In the era of public information disclosure, digital documents have become strategic assets in supporting transparent, accountable, and participatory governance. Effective management of these documents is essential to ensure that public information services are responsive and accessible. However, document classification tasks carried out by Public Information and Documentation Officers (PPID) still rely heavily on manual processes, which are time-consuming, inefficient, and prone to human error. To address this challenge, this study aims to develop an intelligent classification model for public documents using Artificial Intelligence (AI) and Natural Language Processing (NLP), integrated within the Data Lifecycle Management (DLM) framework. The proposed solution was designed using the Design Science Research (DSR) methodology and implemented through Agile development practices. Evaluation was conducted in a simulated laboratory environment that mirrors real-world PPID operations.The developed model leverages transformer-based architectures, particularly BERT (Bidirectional Encoder Representations from Transformers), and is compared against traditional algorithms such as Naive Bayes and K-Nearest Neighbors (KNN). Experimental results show that the BERT model achieves superior performance, with an accuracy of 89%, precision of 0.88, recall of 0.89, and F1-score of 0.88. These metrics confirm that Transformer-based models are highly effective for classifying public documents into categories of information accessibility: available at all times, periodic, immediate, and exempted from disclosure.This research highlights the potential of AI-powered classification to streamline public information services, reduce workload, and enhance compliance with information disclosure laws. The findings support national development priorities such as RPJMN 2025 by contributing to digital transformation in the public sector. The study also provides a replicable framework for other government agencies aiming to implement adaptive and transparent document classification systems.