STR: Seasonal-Trend Decomposition Using Regression

We propose a new method for decomposing seasonal data: seasonal-trend decomposition using regression (STR). Unlike other methods, STR allows multiple and cyclic components, covariates, patterns that may have noninteger periods, seasonality with complex topology. It can be used time series any regular index, including hourly, daily, weekly, monthly, or quarterly data. is competitive existing methods when they exist tackles many more problems than allow. based on regularized optimization so somewhat related to ridge regression. Because it statistical model, we easily compute confidence intervals something not possible most (such as Loess, X-12-ARIMA, SEATS-TRAMO, etc.). Our model implemented in the R package stR, applied by anyone their own