trendbreaker

trendbreaker copied to clipboard

This could be done per model or per CV split for example using the `future` package. Should be loaded optionally.

dirkschumacher

First monitoring system prototype

2

I would suggest setting up a first version of the monitoring system like this: * we always monitor the last 7 days * we train the model on the days...

dirkschumacher

It would be nice to have a metric that takes into account the prediction error and also the proportion of observations within the prediction interval. Since we are mostly interested...

dirkschumacher

See if certain types of boosting models have prediction intervals for regression problems.

dirkschumacher

See if we can also generate good prediction intervals

dirkschumacher

Would be nice to save the pre-trained models for later use. We need to make sure to remove information that is not needed for the prediction. I.e. the model matrix...

dirkschumacher

Fill in some meaningful information and add @thibautjombart

dirkschumacher

dirkschumacher