Model Performance
The linear boosting algorithm was found to be the best model for this forecasting problem, outperforming other algorithms. The use of quantile score as a consistent function for the problem was also highlighted. Probabilistic forecasts, specifically quantile forecasts, provide a richer and more informed forecast, allowing decision-makers to assess the range of possible outcomes.In this clip
From this podcast

Data Skeptic
Water Demand Forecasting
Related Questions