| dc.description.abstract | Climate change and weather variability have become significant global challenges, particularly in predicting atmospheric parameters such as sunshine duration. This parameter plays a crucial role in agriculture, energy, and public health sectors. However, the complexity and uncertainty of meteorological variables such as air temperature, humidity, and rainfall make conventional prediction methods less effective. To address this issue, this study proposes the development of a Mamdani Complex Fuzzy Inference System (M-CFIS), a fuzzy inference model based on complex numbers that comprehensively integrates both amplitude and phase information across the fuzzification, inference, and defuzzification processes. The system was evaluated using temperature, humidity, and rainfall data as input variables to predict sunshine duration. Experimental results show that M-CFIS produces more accurate predictions (with an average of 9.93 hours) compared to the conventional Mamdani method (9.25 hours), while maintaining inference phase stability, indicated by a phase angle of Z approaching zero radians. These findings demonstrate that M-CFIS offers improved accuracy, computational efficiency, and the ability to capture seasonal patterns that are typically unrecognized by conventional fuzzy systems. Furthermore, the model has been successfully implemented in a Python-based application for practical deployment. | en_US |
