M5 Walmart Dataset
A public time series dataset used in the 2020 M5 Forecasting-Accuracy competition. It contains daily sales, price, and promotion information for roughly 30,490 items across seven Walmart stores in the United States from 2011 to 2016, and serves as a benchmark dataset for hierarchical demand forecasting.
Timeline (last 90 days)
On the M5 Walmart dataset, the EventHorizon demand forecasting model achieved a WAPE of 31.03%, while the baseline model's WAPE was 44.4%, achieving roughly a 30% error reduction relative to the baseline.
The developer points out that strong performance on benchmarks does not equal successful deployment in real business environments, and the M5 dataset cannot prove that the same performance can transfer to a specific bakery, flower shop, or other small business.
The M5 Walmart dataset is the public dataset used in the 2020 M5 Forecasting-Accuracy competition, containing time series for about 30,490 items across seven Walmart stores in the United States, covering daily sales from 2011 to 2016, along with price and promotion information.
All Facts (3)
On the M5 Walmart dataset, the EventHorizon demand forecasting model achieved a WAPE of 31.03%, while the baseline model's WAPE was 44.4%, achieving roughly a 30% error reduction relative to the baseline.
50%UnverifiedThe developer points out that strong performance on benchmarks does not equal successful deployment in real business environments, and the M5 dataset cannot prove that the same performance can transfer to a specific bakery, flower shop, or other small business.
50%UnverifiedThe M5 Walmart dataset is the public dataset used in the 2020 M5 Forecasting-Accuracy competition, containing time series for about 30,490 items across seven Walmart stores in the United States, covering daily sales from 2011 to 2016, along with price and promotion information.
50%