Rainfall Prediction in Tegal Regency using ETSFormer

Abstract

Weather is the atmospheric condition at a specific location and time that is variable and constantly changing. Many factors influence weather conditions, one of which is rainfall. Rainfall is a climatic parameter characterized by high variability due to climate anomalies. These anomalies make rainfall prediction very challenging. Specific factors can cause uneven distribution patterns of rainfall between different regions. The exact amount of rainfall that will occur cannot be determined precisely, but predictions or estimations can be made for future rainfall amounts. This study predicts rainfall in Tegal Regency using ETSFormer. Its aim is to provide useful information about future rainfall patterns for the community, especially in Tegal Regency, to facilitate daily activities. The results show that the ETSFormer model effectively predicts rainfall, achieving optimal results with an 8:2 data composition using univariate analysis, yielding the best MSE evaluation metric of 0.002925439039245248 and MAE of 0.036676984280347824.