| Abstrakt: |
Recently the advancement of technology provided numerous ways for predicting the variations of weather from a specific location. In the agricultural field, the success and failure of crop harvesting mainly depend on the amount of rainfall. However, if excess rainfall flows it obtains a challenging role and is not able to predict the accurate rate of rainfall. Elsewhere the early rainfall prediction helped to balance the economic conditions in an agriculture-dominated country like India. Although there have been significant advancements in weather and climate adaptation in recent decades, traditional methods for rainfall prediction remain computationally expensive and complex due to high uncertainty and variability in weather patterns. So to perform an accurate as well as early rainfall prediction the stochastic Bayesian method is proposed that helped to predict the rainfall employed in Indian states such as Uttar Pradesh, Assam, Jharkhand, Tamil Nadu, Andhra Pradesh, etc. Also, the exploration stage is enhanced by tunicate swarm optimization (TSA). The outcomes demonstrate that the scalable stochastic Bayesian approach method was more useful than the existing methods and the accuracy of the rainfall prediction is enhanced by utilizing the crossover-based tunicate swarm algorithm (CTSA). The proposed model is compared in times of MER (%), RMSE (mm), MAPE (%), and MAE (mm). The presented stochastic Bayesian with CTSA achieves 16.23 MER, 4.45 MAPE, 17.96 RMSE, and 13.78 MAE. According to the outcomes, the CTSA algorithm has lower training loss and it proves that the suggested method stochastic Bayesian with CTSA predicts rainfall efficiently. |
Full text from SpringerLink