Shelly Marie Go Bui, Ken Gorro, Gio Angelo Aquino, Mary Jane Sabellano
This study analyzes the Disaster Risk Reduction (DRR) responses on the conducted survey by the National University, Manila, Philippines using k-means clustering. Data preprocessing was implemented to omit irrelevant data for better accuracy. After the data has gone through the algorithm, the study identified labels on the clusters produced such as: attentiveness and alertness, early preparation on upcoming calamities, emergency planning, barangay concern, conduct training seminars/drills, and importance of drainage system. Generally, the results suggest ideas like disaster preparedness, public awareness, and infrastructural improvements. To evaluate k-means results, silhouette Coefficient, an algorithm of intra-cluster distance was used. The result shows an average of approximately 0.0150511564685 which its values is close to 0. It shows that the sample is on or very near to the decision boundary in the nearby clusters. The result of the average silhouette score is acceptable in the distance between each cluster from one another. © 2017 Association for Computing Machinery.
University of San Carlos, Cebu, Philippines
