Skip to content
ML Academy

Introduced on day 22

STL decomposition

A loess-based method that splits a series into trend, seasonal and remainder components; robust options limit the influence of outliers.

In detail

STL (seasonal and trend decomposition using loess) handles any seasonal period, allows the seasonal shape to change over time through the seasonal window, and returns an additive decomposition; take logs first for multiplicative data. Seasonal adjustment subtracts the seasonal component. A multi-period variant (MSTL) exists for daily or hourly data with several seasonal periods.