Modelling of raindrop size distribution using optimized kernel fuzzy c-means clustering algorithm

The Drop Size Distribution (DSD) has been modelled, and the dataset is being fitted using exponential, gamma, and lognormal distribution approaches. The existing Gaussian Mixture Model (GMM) produces the best results, but it ignores or fails to focus on some data points; thus, there is still potential for improvement in the current models. To address these issues, the Optimized Kernel Fuzzy C Means clustering (KFCM) approach is used to effectively cluster data points and predict the drop size distribution. To assess the performance of the proposed model, the Chi-square test is used with rain data from various seasons and types. The results of the proposed model outperformed 11% on seasonal data, whereas the improvement of 30% to 60% is obtained in the case of rain droplets compared to the previous models.