机器学习与语料Free to access
Key facts
Contents & fields
Each dataset consists of three files: ml_{data_name}.csv (temporal graph CSV), ml_{data_name}.npy (edge features), and ml_{data_name}_node.npy (node features). The CSV file contains 5 columns:
- u — user ID
- i — item ID
- ts — timestamp of interaction
- label — label of interaction
- idx — index of interaction
Research uses
Suitable for benchmarking and performance evaluation of temporal graph neural networks, supporting research in dynamic graph representation learning and temporal link prediction.
This card was drafted from the source page; institution, coverage, time span, scale, fields and license are subject to the official page (pending human review).
Keywords
Why this is hard to get on your own
Lack of unified benchmark datasets for evaluating temporal graph neural networks; BenchTemp provides 19 standardized datasets for fair comparison.