Epinions

该数据集是一个包含4,039个节点和88,234条边的无权重和无向图，名为ego-Facebook数据集。为了进行分析，从该数据集中选取了100个节点（编号896至995）及其相应的边，特征数据由576个特征重新调整为24x24的矩阵。此外，该数据集用于评估Conv GT-Net模型在图数据补全方面的性能，包括完成准确度和运行时间，并采用均方误差（MSE）等指标进行评估。数据集的规模包含32个用于测试的图张量和900个用于训练的图张量，总计4,039个节点和88,234条边。所执行的任务是图数据补全。

The ego-Facebook dataset is an unweighted, undirected graph consisting of 4,039 nodes and 88,234 edges. For analytical purposes, 100 nodes (numbered from 896 to 995) and their corresponding edges were selected from this dataset, and the original 576-dimensional feature data was reshaped into a 24×24 matrix. Furthermore, this dataset is utilized to evaluate the performance of the Conv GT-Net model on graph data completion tasks, including completion accuracy and runtime, with metrics such as mean squared error (MSE) employed for assessment. The dataset is split into 900 graph tensors for training and 32 graph tensors for testing, with a total of 4,039 nodes and 88,234 edges. The targeted task of this dataset is graph data completion.