The Performance Evaluation of K-Means and Agglomerative Hierarchical Clustering for Rainfall Patterns and Modelling

The state of rainfall can be identified by using several other climate data. This paper identifies rainfall patterns using other climate data such as temperature, humidity, and sunshine duration. Clusters using K- Means and Agglomerative Hierarchical Clustering (AHC) provide the range of rainfall into three clusters. The purpose of this paper is to analyze the rainfall through the other climate data using different clustering methods. Besides that, it also evaluates the influence of different variables that take effect on rainfall cluster division. This paper evaluates the performance of clustering algorithms using metrics such as silhouette score, Calinski Harabasz score, and Davies Bouldin Score. Silhouette score correlation between rainfall (RR) and duration of sunshine (SS) has the highest score 0.5258 using AHC. Silhouette score correlation between Rainfall (RR) and Average Temperature (Tavg) has the highest 0.5198 using AHC. Finally, the silhouette score correlation between Rainfall (RR) and Average Humidity (RHavg) has the highest score of 0.44 using K-Means.