Predictive asymmetry
#
CausalityTools.CausalityTests.PredictiveAsymmetryTest — Type.
PredictiveAsymmetryTest(predictive_test::CausalityTest)
The parameters for a predictive asymmetry causality test [1].
Mandatory keywords
predictive_test: An instance of a predictive causality test that explicitly uses prediction lags (e.g.VisitationFrequencyTestorTransferOperatorGridTest).
About the prediction lags
The prediction lags in the predictive causality test must consist of n negative integers and n positive integers that are symmetric around zero.
In other words, negative lags must exactly match the positive lags but with opposite sign. The zero lag can be included, but will be ignored, so it is possible to give ranges too.
Examples
test_visitfreq = VisitationFrequencyTest(ηs = [-5, -4, -2, -1, 0, 1, 2, 4, 5])
test_transferoperator = TransferOperatorGridTest(ηs = -3:3)
# Note that `predictive_test` is a *mandatory* keyword.
PredictiveAsymmetryTest(predictive_test = test_visitfreq)
PredictiveAsymmetryTest(predictive_test = test_transferoperator)
References
- Diego, David, Kristian Agasøster Haaga, Jo Brendryen, and Bjarte Hannisdal. A simple test for causal asymmetry in complex systems. In prep.