Multiclass SVM algorithms for wind speed prediction

Abstract

Wind speed prediction has been used in various fields such as Satellite launch, Air traffic control, Weather forecasting etc. Wind speed can be calculated by various atmospheric variables such as temperature, humidity, pressure, wind direction, etc. A number of methods have been proposed by various researchers to predict the wind speed. During the last few years a lot of research has been carried out to forecast the wind speed using several mathematical and biological methods. This paper first reviews various wind speed prediction techniques. It then explores the use of various multiclass Support Vector Machine (SVM) algorithms for wind speed prediction. It has been shown that multiclass Directed Acyclic Graph based Support Vector Machine algorithm produces better results than other multiclass SVM algorithms.