Home › Stats-TimeSeries › Seasonality

SeasonalComponents

A series split into its trend, its seasonal pattern and what neither explains.

public sealed class SeasonalComponents

Properties — Trend is the centred, or trailing, moving average. Seasonal is the average detrended value at each phase, centred, tiled over the series. Residual is the series less the trend and the seasonal component, or divided by them under the multiplicative model.

Example — the additive components add back to the series wherever the trend is defined.

using Lodestar.Stats.TimeSeries;

double[] quarterly = [10.0, 14.0, 8.0, 12.0, 11.0, 15.0, 9.0, 13.0, 12.0, 16.0, 10.0, 14.0];

SeasonalComponents parts = SeasonalDecomposition.Decompose(quarterly, period: 4);

double rebuilt = Math.Round(parts.Trend[5] + parts.Seasonal[5] + parts.Residual[5], 4);  // => 15
int length = parts.Seasonal.Count;                                                       // => 12

Remarks — there is no public constructor. The components are what SeasonalDecomposition.Decompose returns. The trend and the residual are NaN where the moving average has no full window, unless SeasonalDecompositionOptions.ExtrapolateTrend filled them.

A class rather than a record, for the reason AutocorrelationResult gives.

Applies to — net10.0, netstandard2.0.

See also — SeasonalDecomposition.Decompose.