recursive predictions
Can you make recursive predictions?
Hi,
I never tried and I'm not completely sure what you mean. But, let's say that at inference time the forecasting model progressively generates new predictions (with an associated confidence interval) by using as input the previous predictions. Then yes, it is possible to adjust such confidence intervals with the offsets estimated with conformal prediction on the calibration set.
I hope this answers your question.
Thank you very much for your answers. I apologize for disturbing you again. I would like to consult you on another issue. If I want to directly use the TCN model for ensemble conformalized quantile regression, I would like to consult on the specific steps. Should I directly call the EnCQR method, or do I need to first use the TCN model for training and prediction, that is, call the corresponding fit method and transform method, and then call the EnCQR method?