Wind Speed Prediction in Jerusalem Using Machine Learning Algorithms: A Case Study of Using ANFIS and KNNR

Abstract

Wind energy is acknowledged for its status as a renewable energy source that offers several advantages, including its low cost of electricity generation, abundant availability, high efficiency, and minimal environmental impact. The prediction of wind speed using machine learning algorithms is crucial for various applications, such as wind energy planning and urban development. This paper presents a case study on wind speed prediction in Palestine Jerusalem city using the Adaptive Neuro-Fuzzy Inference System (ANFIS) and K-Nearest Neighbors Regression (KNNR) algorithms. The study evaluates their performance using multiple metrics, including root mean square (RMSE), bias, and coefficient of determination R<sup>2</sup>. ANFIS demonstrates good accuracy with lower RMSE (0.196) and minimal bias (0.0003). However, there is room for improvement in capturing overall variability (R<sup>2</sup> = 0.15). In contrast, KNNR exhibits a higher R<sup>2</sup> (0.4093), indicating a better fit, but with a higher RMSE (1.4209). These results demonstrated the potential of machine learning algorithms in wind speed prediction, which can lead to optimize the wind energy generation at specific site, and reducing the cost of energy production. This study provides insights into the applicability of ANFIS and KNNR in wind speed prediction for Jerusalem and suggests future research directions. The outcomes have practical implications for wind energy planning, urban development, and environmental assessments in similar regions.