<![CDATA[Weather refers to atmospheric conditions such as temperature, humidity, air pressure, wind speed, and rainfall, all of which influence human activities. Rainfall is particularly important due to its impact on agriculture and water resource management. This study classifies regions on Java Island based on rainfall patterns using the Fuzzy C-Means algorithm. Rainfall variations are influenced by geographical, topographical, and climatic factors, requiring methods that can capture spatial and temporal changes. Fuzzy C-Means was selected for its ability to manage data uncertainty and overlapping clusters. To measure rainfall pattern similarity between regions, the Dynamic Time Warping (DTW) method was applied. Since DTW is a non-Euclidean metric and incompatible with Fuzzy C-Means, the Multidimensional Scaling (MDS) method was used to convert DTW distance matrices into Euclidean feature vectors. The study used secondary daily rainfall data from NASA (2021–2024). Clustering performance was evaluated using the Silhouette Coefficient, yielding a value of 0.413184, indicating good compactness and separation. Results identified three clusters: low rainfall (Cluster 0), moderate rainfall (Cluster 1), and high rainfall (Cluster 2). ANOVA results confirmed significant differences in average rainfall between clusters, with Tukey HSD tests showing Cluster 2 significantly differs from Clusters 0 and 1, while Clusters 0 and 1 are not significantly different. These findings demonstrate that combining DTW, MDS, and Fuzzy C-Means effectively identifies temporal rainfall patterns and produces statistically meaningful clustering. The spatial distribution of each cluster is visualized using GeoJSON and a database for clearer interpretation.]]>
