Air Pollution Index (API) Analysis at Jakarta in 2019-2020 using Fuzzy C-Means and Gaussian Mixture Model

This study aims to compare the Air Pollution Index (API) clustering between fuzzy c-means (FCM) with gaussian mixture model. This study used air quality data on each parameter in 2019-2020 from five monitoring stations, that is Bundaran HI (DKI1), Kelapa Gading (DKI2), Jagakarsa (DKI3), Lubang Buaya (DKI4), and Kebon Jeruk (DKI5). Determination of the optimum cluster number on Fuzzy C-Means based on Partition Coefficient (PC), Classification Entropy (CE), Separation Index (SI), Silhouette Index, and Effectiveness. The optimum cluster number in the Gaussian Mixture Model is based on BIC and Silhouette Index values. Almost all Silhouette values on Fuzzy C-Means are more significant than the Silhouette Gaussian Mixture Model. Fuzzy C-Means is more suitable for clustering Jakarta Air Pollution Index (API) than The Gaussian Mixture Model method.