<![CDATA[Resource-speaker management at LPP RRI Manado has so far depended on manual, non-integrated procedures, which has made attendance monitoring and objective activity evaluation difficult. This study builds a web-based Resource Speaker Management System that embeds the K-Means Clustering algorithm to examine speaker activity. The application was developed with PHP, MySQL, HTML, CSS, and JavaScript following the Waterfall model. It was tested on 50 resource speakers covering 52 scheduling records and 33 confirmed attendances, giving an average consistency of 65%. Running K-Means with K=3 separated the speakers into three groups: Highly Active (2%, 1 person), Moderately Active (64%, 32 people), and Less Active (34%, 17 people). The converged normalised centroids of the three clusters were (1; 1; 0.50), (0; 1; 1.00), and (0.059; 0; 0), respectively. The system simplifies scheduling administration and photo-evidenced attendance recording, and yields an objective grouping that can underpin speaker-performance evaluation for the management of LPP RRI Manado.]]>
Copyrights © 2026
