@krzysztofsztompka/garch
    Preparing search index...

    Type Alias ValidationOptions

    walkForwardValidation's specification: the model, the horizon, the windows and the baseline.

    type ValidationOptions = {
        baseline: Baseline;
        dist?: Distribution;
        exogenous?: ArrayLike<number>;
        horizon: number;
        mean?: MeanModel;
        minTrain?: number;
        model: GarchModel;
        step?: number;
        targetVariance?: boolean;
    }
    Index
    baseline dist? exogenous? horizon mean? minTrain? model step? targetVariance?
    baseline: Baseline

    What the model must beat: naive — the trailing horizon realised variance (the random-walk vol forecast); garch — the symmetric model, for asking whether an asymmetric term earns its keep.

    dist?: Distribution
    exogenous?: ArrayLike<number>

    GARCH-X: x aligned with the returns (garch and gjr); see FitOptions.exogenous. Forecasts hold it at its last value.

    horizon: number
    mean?: MeanModel

    A fitted constant (default) or zero; see FitOptions.mean.

    minTrain?: number
    model: GarchModel
    step?: number
    targetVariance?: boolean

    Variance targeting (garch and gjr); see FitOptions.targetVariance.

    Settings
    Member Visibility

    On This Page
    Properties