Time Series Clustering Analysis Using Dynamic Time Warping Technique of Daily Rainfall in Bengkulu Province

This title has been presented on Thursday, December 14, 2023 at 13.50-14.00 GMT+7.

Abstract

This title has been presented on Thursday, December 14, 2023 at 13.50-14.00 GMT+7.

Rainfall is a significant problem. The Meteorology, Climatology and Geophysics Agency's findings indicate that the state of climate change is critical. High rainfall can have an adverse effect on people's daily lives by contributing to numerous floods, landslides, and other disasters. As a result, this researcher will use the time series clustering approach to analyze the mapping of rainfall in Bengkulu Province from 2001 to 2020. In this work, the size of similarity is determined using the Euclidean and Dynamic Time Warping (DTW) approaches. then create a dendrogram using the hierarchical technique. The cophenetic correlation coefficient's ideal equivalent. Re-clustering using k-Means by using the cluster center from the calculation of the hierarchical technique and choosing the best. The results obtained were that there were 2 clusters with the number of members in cluster 1 being 1 rain station, namely manna and cluster 2 having 14 rain stations. Cluster 1 is included in the low rainfall category. Cluster 2 is a cluster with high rainfall category.