WOLFRAM

Wolfram Language & System Documentation Center

Fit Time Series Processes to Data

Fit an autoregressive process of order 6 to the following data:
Wolfram Language code: ListLinePlot[data = TemporalData[Automatic, {{{0., 0.09, -0.05999999999999997, -0.010000000000000009, -0.040000000000000036, 0.010000000000000009, 0.040000000000000036, -0.07, 0.07, 0.10999999999999999, -0.22999999999999998, 0.12, -0.050000000000000044, -0.0 ... 0., 0.13, 0.07000000000000006, -0.07000000000000006, 0., 0.13, -0.030000000000000027, -0.14999999999999997, 0., 0.12999999999999995, -0.05999999999999994}}, {{0, 131, 131/130}}, 1, {"Discrete", 1}, {"Discrete", 1}, 1, {}}, False, 9.]]
Wolfram Language code: eproc = EstimatedProcess[data, ARProcess[6]]
The partial correlation of the fitted model and data agree well:
Wolfram Language code: Labeled[Show[ListPlot[TemporalData[PartialCorrelationFunction[#, {30}]& /@ {eproc, data}], Filling -> {1 -> {2}}, FillingStyle -> Thick, PlotStyle -> PointSize[Large], PlotRange -> {-.4, .2}, PlotLegends -> Placed[Style[#, Bold, FontFamily -> "Verdana"]& /@ {"Data", "AR[6]"}, Bottom]], Graphics[{Dashed, Thick, Line[{{0, #}, {35, #}}]}]& /@ (Quantile[NormalDistribution[], {(1 - .95) / 2, 1 - (1 - .95) / 2}] / Sqrt[data["PathLengths"][[1]]]), ImageSize -> 400], Column[{Style["Partial Auto‐Correlation Plot", Bold, FontSize -> 18, FontFamily -> "Verdana"], Style["95% Confidence", Bold, FontSize -> 16, FontColor -> Gray, FontFamily -> "Verdana"]}], {{Top, Left}}]