Accurate local very short-term temperature prediction based on synoptic situation Support Vector Regression banks

In this paper we present a novel system for addressing problems of local very short term (up to a time prediction horizon of 6 h) temperature prediction based on Support Vector Regression algorithms (SVMr). Specifically, we construct SVMr banks based on the synoptic situation for each prediction period, incorporated by means of the well-known Hess–Brezowsky classification (HBC). We show how this SVMr bank structure obtains very good results in a real problem of short-term temperature prediction at Barcelona-El Prat International Airport (Spain), obtaining an average RMSE of 1.34 °C in 6 hour horizon prediction. Comparison with alternative neural techniques have been carried out in order to show the effectiveness of the proposed technique, and how the inclusion of the HBC classification is also able to improve the performance of these alternative neural algorithms in the problem.

► Banks of neural algorithms improve the prediction of short-term local temperature.
► Synoptic situation is included in the prediction system.
► Support Vector Regression banks obtained the best results in this problem.