Forecasting metrics: normalization and aggregation Calculating metrics for forecasting model evaluation requires some special considerations when the data contains multiple time series. There are two natural choices for aggregating metrics over multiple series: A macro average wherein the evaluation metrics fr...
followed by an illustration of the graph representation, structure similarity profile, sparse matrix representation, and the procedure to update the similarity score. In addition, an explanation is provided for the evaluation metrics used
Forecasting metrics: normalization and aggregationCalculating metrics for forecasting model evaluation requires some special considerations when the data contains multiple time series. There are two natural choices for aggregating metrics over multiple series:...