Cloudy1225/HTAG

该数据集是多尺度异质文本属性图数据集，涵盖了多个领域，包括电影合作、社区问答、学术、书籍出版和专利申请等。数据集规模从小型（24K节点，104K边）到大型（5.6M节点，29.8M边）不等，适用于测试计算密集型算法和开发可扩展模型。数据集提供了自动化的评估管道，确保评估的可重复性，并采用了基于时间的数据分割，以提供更真实和有效的评估。此外，数据集的构建代码已开源，支持研究人员构建更大更复杂的异质文本属性图数据集。

This dataset is a multi-scale heterogeneous text-attributed graph dataset that spans multiple domains, including movie collaboration, community question answering, academic, book publication, and patent application. The dataset scales range from small (24K nodes, 104K edges) to large (5.6M nodes, 29.8M edges), making it suitable for testing computationally intensive algorithms and developing scalable models. The dataset provides an automated evaluation pipeline to ensure reproducibility and employs time-based data splits for more realistic and meaningful evaluation. Additionally, the dataset construction code is open-source, enabling researchers to build larger and more complex heterogeneous text-attributed graph datasets.