@krzysztofsztompka/garch
    Preparing search index...

    Type Alias DccFit

    A two-stage DCC-GARCH fit: each asset's fit, the DCC(1,1) dynamics and every day's correlation.

    type DccFit = {
        a: number;
        assets: readonly GarchFit[];
        b: number;
        converged: boolean;
        correlation: Float64Array[];
        correlationLogLikelihood: number;
        correlationTarget: number[][];
        lastQ: number[][];
        lastStandardized: number[];
        logLikelihood: number;
    }
    Index
    a assets b converged correlation correlationLogLikelihood correlationTarget lastQ lastStandardized logLikelihood
    a: number

    The weight of yesterday's shock in Q_t.

    assets: readonly GarchFit[]

    Stage 1: each asset's fit, in the order given.

    b: number

    The weight of yesterday's Q in Q_t; a + b < 1.

    converged: boolean

    Stage 2 converged, and so did every asset's fit.

    correlation: Float64Array[]

    R_t for every day, each N×N flattened row by row.

    correlationLogLikelihood: number

    Stage 2's log-likelihood: the correlation part.

    correlationTarget: number[][]

    Q̄, the standardised residuals' second moment the dynamics revert to (N×N).

    lastQ: number[][]

    Q_T, where the correlation recursion stands after the last day (N×N).

    lastStandardized: number[]

    z_T, the last day's standardised residuals.

    logLikelihood: number

    Stage 1's log-likelihoods plus stage 2's — the joint one when every asset is normal.

    Settings
    Member Visibility

    On This Page
    Properties