Iterative gradient ascent expected maximization clustering for weather forecasting

Abstract

Weather forecasting is a significant process to be solved as it discovers future atmosphere for a given location. Few clustering techniques were intended in order to group similar weather data for predicting weather conditions. However, the clustering accuracy of the existing technique was not effectual when taking big dataset as input. In order to solve this limitation, a Iterative Gradient Ascent Expected Maximization Clustering (IGAEM) Model is proposed. The IGAEM Model predicts the future weather conditions with higher clustering accuracy and minimal time. In IGAEM Model; after selecting the relevant features, IGAEM Model applied Iterative Gradient Ascent Expected Maximization Clustering (IGAEMC) to accurately group the weather data into diverse clusters with lower amount of time utilization. Thus, IGAEM Model significantly increases the performance of weather forecasting as compared to existing works. The IGAEM Model conducts experimental evaluation using factors such as clustering accuracy clustering time and false positive rate with respect to a number of features and weather data from Atlantic hurricane database. The experimental results depict that the IGAEM Technique is able to enhance the clustering accuracy and reduce the clustering time of weather prediction when compared to state-of-the-art-works.