首页|Researcher at University of Limpopo Zeroes in on Machine Learning(Forecasting S hort- and Long-Term Wind Speed in LimpopoProvince Using Machine Learning and Ex treme Value Theory)
Researcher at University of Limpopo Zeroes in on Machine Learning(Forecasting S hort- and Long-Term Wind Speed in LimpopoProvince Using Machine Learning and Ex treme Value Theory)
扫码查看
点击上方二维码区域，可以放大扫码查看
原文链接
NETL
NSTL
By a News Reporter-Staff News Editor at Robotics & Machine Learning DailyNews Daily News – Researchers detail new data in artific ial intelligence. According to news reportingfrom Sovenga, South Africa, by New sRx journalists, research stated, “This study investigates wind speedprediction using advanced machine learning techniques, comparing the performance of Vanill a long shorttermmemory (LSTM) and convolutional neural network (CNN) models, a longside the application ofextreme value theory (EVT) using the r-largest order generalised extreme value distribution (GEVDr).Over the past couple of decades , the academic literature has transitioned from conventional statistical time se ries models to embracing EVT and machine learning algorithms for the modelling o f environmentalvariables.”Our news reporters obtained a quote from the research from University of Limpopo : “This study addsvalue to the literature and knowledge of modelling wind speed using both EVT and machine learning. Theprimary aim of this study is to foreca st wind speed in the Limpopo province of South Africa to showcase thedependabil ity and potential of wind power generation. The application of CNN showcased con siderablepredictive accuracy compared to the Vanilla LSTM, achieving 88.66% accuracy with monthly time steps.The CNN predictions for the next five years, i n m/s, were 9.91 (2024), 7.64 (2025), 7.81 (2026), 7.13(2027), and 9.59 (2028), slightly outperforming the Vanilla LSTM, which predicted 9.43 (2024), 7.75 (2025), 7.85 (2026), 6.87 (2027), and 9.43 (2028). This highlights CNN’s superior ab ility to capture complexpatterns in wind speed dynamics over time. Concurrently , the analysis of the GEVDr across various orderstatistics identified GEVDr=2 a s the optimal model, supported by its favourable evaluation metrics interms of Akaike information criteria (AIC) and Bayesian information criteria (BIC). The 3 00-year returnlevel for GEVDr=2 was found to be 22.89 m/s, indicating a rare wi nd speed event.”
University of LimpopoSovengaSouth Af ricaAfricaCyborgsEmerging TechnologiesMachine Learning
NETL
NSTL
