This block is a continuation of this previous block. Both experiments how to detect if a timeline has a seasonality component, using a correlogram. This block explains why detrending the time serie before computing the correlogram is a must have.
Usages :
- same usages as in this previous block
- in the correlogram, detrending the time serie before computing the correlogram allows to detect very small seasonnality order of magnitude; detrending the time serie implies that coefficients of correlation for each lag no longer reflect any trend, and thus only reflect the seasonality component;
- while using detrended time serie, increasing/decreasing trend has no longer any impact on the correlogram; chery on the cake, seasons are easier to detect.
Notes:
- the previous block experiments season detection without detrending
- another block experiments autocorrelation
- another block experiments time series correlation
- another block deals with the impact of seasonality when computing the trend of a timeline
Acknowledgments: