Devni Prima Sari, Media Rosha
There are several efforts to mitigate earthquake disasters. One is to do regional planning following earthquake hazard studies, such as using seismic hazard analysis. The hazard that occurs later is in the form of ground acceleration in the bedrock, which is defined by the PGA (Peak Ground Acceleration) value. PGA is a continuous variable, so variable discretization is carried out. This process is claimed to save memory usage, improving knowledge representation because the data is simpler to understand. The application of mining techniques makes the performance faster and perfect. K-Means is a data grouping method that performs an unsupervised modeling process. In addition to grouping, K-means can also be used to discretize variables. However, sometimes the K-Means algorithm does not give the best results because it is sensitive to outliers. Outliers are points that differ from other data points. Based on these findings, we tried to discretize PGA with the K-medoids method, where the discretization validation level for the PGA variable was 98% based on the silhouette coefficient data. © 2023 Author(s).
Department of Mathematics, Universitas Negeri Padang, Padang, Indonesia
