GrangerFunc: GrangerFunc

In VLTimeCausality: Variable-Lag Time Series Causality Inference Framework

Description

GrangerFunc is a Granger Causality function. It tests whether X Granger-causes Y.

Usage

GrangerFunc(
  Y,
  X,
  maxLag = 1,
  alpha = 0.05,
  autoLagflag = TRUE,
  gamma = 0.5,
  family = gaussian
)

Arguments

Y	is a numerical time series of effect
X	is a numerical time series of cause
maxLag	is a maximum possible time delay. The default is 1.
alpha	is a significance level of F-test to determine whether X Granger-causes Y. The default is 0.05.
autoLagflag	is a flag for enabling the automatic lag inference function. The default is true. If it is set to be true, then maxLag is set automatically using cross-correlation. Otherwise, if it is set to be false, then the function takes the maxLag value to infer Granger causality.
gamma	is a parameter to determine whether X Granger-causes Y using BIC difference ratio.
family	is a parameter of family of function for Generalized Linear Models function (glm). The default is gaussian.

Value

This function returns of whether X Granger-causes Y.

ftest	F-statistic of Granger causality.
p.val	A p-value from F-test.
BIC_H0	Bayesian Information Criterion (BIC) derived from Y regressing on Y past.
BIC_H1	Bayesian Information Criterion (BIC) derived from Y regressing on Y,X past.
XgCsY	The flag is true if X Granger-causes Y using BIC difference ratio where BICDiffRatio >= gamma.
XgCsY_ftest	The flag is true if X Granger-causes Y using F-test where p.val>=alpha.
XgCsY_BIC	The flag is true if X Granger-causes Y using BIC where BIC_H0>=BIC_H1.
maxLag	A maximum possible time delay.
H0	glm object of Y regressing on Y past.
H1	glm object of Y regressing on Y,X past.
BICDiffRatio	Bayesian Information Criterion difference ratio: (BIC_H0-BIC_H1)/BIC_H0.

Examples

# Generate simulation data
TS <- SimpleSimulationVLtimeseries()
# Run the function
out<-GrangerFunc(Y=TS$Y,X=TS$X)

VLTimeCausality documentation built on Jan. 24, 2022, 5:07 p.m.