Explore

Q: How can I perform hierarchical reconciliation: of a variety of base forecast models with the MinT algo (ideally), with nonnegative reconciled forecast outputs (necessarily), which are distributed (necessarily, either closed form or simulated), within the R tidyverts/fable framework (ideally)? Here are a couple options I tried. Unfortunately, both fall short of my needs: fabletools::min_trace(method='mint_shrink'): I love that this has the Unfortunately, this does...

The Bureau of Meteorology’s latest long-range forecast of Australia’s climate outlook predicts potentially record warmer May to July days and nights for most of Australia. This year’s unusually warm winter means maximum and minimum temperatures are very likely to be above median for most of the continent. Except for parts of northern NT and Queensland, […]