refunders / refund

Regression with functional data
39 stars 23 forks source link

predict.pffr fails for concurrent functional effects #87

Closed fabian-s closed 1 year ago

fabian-s commented 5 years ago

.. the conditionals starting in https://github.com/refunders/refund/blob/fa0c2c03f23fb8a4ecffbf0dd79a1a80cc8cf518/R/pffr-methods.R#L109 don't check for matrix covariates, so only their first few values get repeated in gamdata[[cov]] <- drop(newdata[[cov]])[rep(1:nobs, each=nyindex)]