Time series analysis and forecast, Smoothing

Examples: Time series analysis

Smoothing

The 'Caterpillar'-SSA is an appropriate technique for smoothing time series. Natural decomposition of the time series is constructed on the base of the unique parameter (the window length). Grouping slowly changing components, you can separate the low-frequency components from the high-frequency components of the time series.

Figure 1.3: Tree rings: smoothing result

Figure 1.4: Tree rings: residuals

Data: "Tree rings"
Description: tree ring indices, Douglas fir, Snake river basin, U.S., annual, from 1282 to 1950
Source: Time Series Data Library maintained by Rob Hyndman
Parameters: window length = 120
Figure 1.3: the time series component reconstructed from the leading 7 eigentriples (red line) and the initial time series
Figure 1.4: residuals of the reconstruction from the leading 7 eigentriples

Previous example | Next example