This function fits a Gaussian Process model to a data.frame and then generates predictions from the fitted model.
Usage
fit_predict_gp(
df,
t_steps = 1000,
complete_sigma = FALSE,
prepare_dataframe = FALSE,
estimate_derivatives = FALSE,
delete = TRUE,
pred_for_t = NULL
)Arguments
- df
A data.frame with the data that is to be modelled. Must have timepoint, timepoint_n, curve_id, and od fields.
- t_steps
the number of timepoints within the range od df$timepoints to predict
- complete_sigma
A logical indicating whether the complete covariance matrix should be returned. This is necessary if you want to re-sample the posterior.
- prepare_dataframe
A logical indicating whether a data.frame with the mean data should be prepared.
- estimate_derivatives
A logical indicating whether the empirical derivatives should be computed.
- delete
A logical indicating whether the GP model should be deleted from memory FALSE.
- pred_for_t
A numeric value inidicating a specific timepoint t for which OD should be predicted