Wind Prediction Based on General Regression Neural Network

Abstract

This study adopts the general regression neural network (GRNN) to predict wind speeds. The training data sets are the real wind speeds obtained from CKS International Airport. The 5 days (120 hours) of the three year from 2006 to 2008 is selected as an example to appraise the prediction performance by using GRNN. Comparing to the traditional linear time-series-based model, the superiority of GRNN method to wind prediction can be valid.