<![CDATA[The Indonesian Ulema Council (MUI) of Riau Province faces challenges in objectively evaluating dakwah (Islamic preaching) as its existing Peta Dakwah Cerdas (PDC) system lacks a feature to analyze congregant feedback. This study aims to design and implement a topic modeling model to identify the main hidden themes within congregant opinions. The study utilized 2,581 comments collected from the MUI Riau Smart Evaluation System. The methodology involved text preprocessing, Term Frequency-Inverse Document Frequency (TF-IDF) word weighting, and topic modeling using the Non-Negative Matrix Factorization (NMF) algorithm. Toward determine the optimal number of topics (k), the model was evaluated using Coherence Score to measure semantic readability and Silhouette Score to measure the resulting topic separation. The experiment identified two topics (k=2) as the best configuration achieving a high Coherence Score of 0.7023 and a Silhouette Score of 0.0163. The two main topics formed represent (1) Prayers and Greetings for the Preacher, and (2) Congregant Participation and Appreciation for the Dakwah. The application of NMF proved effective in identifying thematic patterns in congregant opinions and can serve as a foundation for MUI Riau to develop a real-time Islamic preaching evaluation system.]]>
Copyrights © 2025
