| # Dataset information: the entry point for dataset configuration, including file locations for each modality. | |
| dataset_id: M5 | |
| dataset_name: M5 Forecasting Accuracy | |
| domain: retail_demand_forecasting | |
| sample_defination: | |
| sample_type: item-store sales episode | |
| sample_id_graph: episode_id | |
| description: each sample corresponds to one item sold at one Walmart store | |
| modalities: | |
| # Available feature modalities. | |
| meta_features: | |
| required: true | |
| file_name: meta.csv | |
| node_level_task_graphs: | |
| available: true | |
| graph_ids: | |
| - item_graph | |
| time_series: | |
| available: true | |
| time_series_ids: | |
| - calendar_month | |
| - calendar_wday | |
| - calendar_year | |
| - sell_price | |
| - snap_CA | |
| - snap_TX | |
| - snap_WI | |
| - unit_sales | |
| tasks: | |
| # Supported task types, for example meta_feature prediction, time series forecasting, and ranking. | |
| - item_category_classification | |
| - episode_unit_sales_forecasting | |
| # Data source information. | |
| license: | |
| data_license: Kaggle competition data terms | |
| redistribution_allowed: source-dependent | |
| source: | |
| name: M5 Forecasting - Accuracy | |
| url: https://www.kaggle.com/competitions/m5-forecasting-accuracy | |