Analysis of Changes in Weather Elements Clustering in Pagaralam in 2023 and 2024 Using K-Means and Average Linkage Clustering Methods Sri Indra Maiyanti (a), Irmeilyana (a*), Putri Nilam Cayo (a), Ngudiantoro (a)
a) Department of Mathematics, Faculty of Mathematics and Natural Sciences, University of Sriwijaya
Jl. Raya Palembang-Prabumulih Km. 32, Indralaya, South Sumatra Province, Indonesia
*irmeilyana[at]unsri.ac.id
Abstract
Climate is a combination of weather elements over a long period of time. Global climate change can impact the productivity of agricultural crops, including coffee plantations in Pagaralam, Indonesia. The purpose of this study is to cluster the time of occurrence based on weather elements data in Pagaralam in 2023 and 2024 using K-means and average linkage clustering methods. The data matrix consists of 53 weeks and 15 weather element variables. Based on the biplot and Silhouette Index results of both data matrices, the K values used were 3 and 4. The K-means clustering results on the 2024 data matrix indicate that a cluster of the majority of weeks is characterized by higher minimum temperature, dew point, humidity, precipitation cover, and cloud cover. Meanwhile, in the 2023 data matrix, a cluster of the majority of weeks is characterized by high levels of dew, humidity, and cloud cover. The average linkage clustering results are also similar to the K-means clustering results.
Keywords: Average linkage- cluster- K-means clustering- Pagaralam- weather elements
