<![CDATA[Indonesia, as a tropical country with high rainfall, heavily relies on accurate rainfall predictions for various critical purposes, including water resource management and extreme weather impact mitigation. One commonly used method is the Tsukamoto Fuzzy Inference System (FIS). However, implementing the Tsukamoto FIS often leads to high error rates. This is attributed to the difficulty in determining the boundaries of fuzzy variable membership functions. To address this issue, this research proposes an innovative approach by optimizing the boundaries of fuzzy membership functions using Genetic Algorithms (GA). The study resulted in a 49.02% reduction in the error rate, decreasing from 76.82% to 27.8%. This method significantly enhances rainfall prediction accuracy and contributes to the advancement of more sophisticated prediction methods. The optimization method proposed in this study also holds potential for application across various atmospheric science contexts.]]>
Copyrights © 2024
