Seasonal outliers in time series

Abstract

In the analysis of time series, it is frequent to classify perturbations as Additive Outliers (AO) , Innovative Outliers (10), Level Shift (LS) outliers or Transitory Change (TC) outliers. When a time series with a clear seasonal behaviour is considered, this classification may be too restrictive since none of the four outlier types is adequate to model changes in the seasonal pattern of the series. In this paper, a new outlier type, the Seasonal level Shift (SLS), is introduced in order to complete the usual classification. The iterative procedure for the detection of outliers in Chen and Liu (1993) is extended to detect SLS outliers. We use simulations and real examples to assess the properties of the new type of outlier.