ANALYSIS OF LIBRARY VISITOR GROUPING THROUGH MASK USAGE IDENTIFICATION IN XIN ZHONG LIBRARY WITH ORANGE DATA MINING APPLICATION

Abstrak

Abstract
The application of data mining in libraries plays a crucial role in supporting data management and monitoring health protocols, especially during the pandemic. A key challenge faced by librarians is effectively monitoring visitors' mask usage compliance. This study aims to analyze visitors' facial images at the library using the Orange Data Mining application, enabling librarians to identify whether visitors are wearing masks. The approach involves collecting random facial images of visitors, preprocessing the data for standardization of size and resolution, extracting features using the Inception V3 model, and conducting hierarchical clustering analysis with the Manhattan metric. The clustering results are visualized in a dendrogram, helping to group the data. The findings show that the dendrogram clearly differentiates between visitors with masks and those without. This visualization provides librarians with an effective tool for monitoring areas of the library that require more strict health protocol supervision. The study concludes that the Orange Data Mining application offers a practical solution for libraries to monitor compliance with health protocols. By utilizing data mining techniques, libraries can enhance visitor safety and comfort. Further research is suggested to expand the dataset and explore other methods to improve analysis accuracy.