search

MatthieuStigler / tsDyn

tsDyn
tsdyn.googlecode.com
34 stars 20 forks source link

Heteroskedasticity #9

Closed Tsumbu closed 3 years ago

Tsumbu commented 5 years ago

Dear Mr Matthieu Stigler, I am using the SETAR and the LSTAR model to model and forecast return. I want to ask if there is a way to account for heteroskedasticity in the tsDyn package so that I can have correct inference.

I am particularly interested in accounting for heteroskedasticity in the setarTest(), lstar() and in the setar() function.

MatthieuStigler commented 5 years ago

Hi

There is no methods for now to allow HC covariance in tsDyn. For SETAR models, you can just extract the model-matrix from the SETAR output and re-run it using lm() and sandwhich. For LSTAR, this is more tricky, I am not sure how to derive the relevant expressions.

MatthieuStigler commented 3 years ago

Closing this issue as inactive for a while.