Short-Term Probabilistic Forecasting Method for Wind Speed Combining Long Short-Term Memory and Gaussian Mixture Model

Wind speed forecasting is advantageous in reducing wind-induced accidents or disasters and increasing the capture of wind power. Accordingly, this process has been a focus research field engineering. However, because chaotic random nature, its inevitably includes errors. Consequently, specifying appropriate method to obtain accurate results difficult. The probabilistic considerable relevance short-term it provides both predicted value error distribution. This study proposes for speeds based on Gaussian mixture model long memory. precision proposed evaluated by prediction intervals (i.e., interval coverage probability, normalized average width, width-based criterion) using 29 monitored datasets. effects characteristics were further studied. Results show that effective obtaining probability distribution speeds, forecast are highly accurate. mainly influenced difference standard deviation.