The effect of different state sizes on Mycielski approach for wind speed prediction

Abstract

Short-term wind speed prediction is very useful for efficient utilization of wind power. Mycielski algorithm is a novel and effective approach alternative to statistical or artificial intelligence based models for wind speed prediction. In order to apply the Mycielski algorithm, the wind speed values are converted to wind states. This paper investigates the effect of different wind speed state sizes on the Mycielski approach for wind speed prediction. The data sets from three different wind farms are used to test the variation of different state sizes for the Mycielski approach. The comparison results and F-test statistics show that the moderate small state size has a significant benefit to predict the wind speed accurately.