Community-imbalanced graph dataset

The dataset are served for community-imbalanced graph sampling algorithm performance experiments. In the algorithm performance experiment, we selected 30 graph datasets, 15 of which were derived from real-world graph datasets (https://snap.stanford.edu/data/), and 15 were adapted from real-world datasets or simulated datasets. The simulated datasets can address the problems of incomplete coverage of imbalance degree of community and uncertain community detection results in real-world datasets. These 30 datasets encompass varying node and edge scales, different numbers of communities, and varying degrees of community imbalance. Moreover, these datasets encompass various application domains. Details about these graph datasets were provided in dataset description.

本数据集旨在服务于社区不平衡图采样算法的性能实验。在算法性能实验中，我们选取了30个图数据集，其中15个源自现实世界的图数据集（https://snap.stanford.edu/data/），另15个则改编自现实世界数据集或模拟数据集。模拟数据集能够解决现实世界数据集中社区不平衡程度覆盖不全面和社区检测结果不确定的问题。这30个数据集涵盖了不同的节点和边规模、不同的社区数量以及不同程度的社区不平衡。此外，这些数据集还覆盖了多个应用领域。有关这些图数据集的详细信息已包含在数据集描述中。