Resumen de M-Estimación aplicada a suavización exponencial con observaciones faltantes y otras anómalas

Diego Bote Barco, Asunción Rubio de Juan

• The paper is devoted to the study of exponential smoothing when some special characteristics are present in data, i.e., missing observations and outliers. Classical exponential smoothing is not appropriate in these situations because it gives bad results or cannot be applied at all. The improvement is based on M-estimation principle which is robust when outliers are present in data including some modifications for missing observations.