未明确提及

本文介绍了一种新的基准图生成方法，用于测试有向和加权图中的社区检测算法，特别是考虑了节点可能属于多个社区的情况。该方法扩展了作者先前提出的无向和无权重基准图的概念，引入了节点度和社区大小的异质分布。数据集的创建过程涉及算法生成具有内置社区结构的网络，并考虑了网络中链接的方向和权重。该数据集适用于评估社区检测算法的性能，特别是在处理真实世界网络中常见的复杂社区结构时。

This paper presents a novel benchmark graph generation method for testing community detection algorithms on directed and weighted graphs, with particular consideration of the scenario where nodes may belong to multiple communities. This method extends the concept of undirected and unweighted benchmark graphs previously proposed by the authors, and introduces heterogeneous distributions of node degrees and community sizes. The dataset creation process involves generating networks with built-in community structures via algorithms, while taking into account the direction and weight of links in the network. This dataset is suitable for evaluating the performance of community detection algorithms, especially when dealing with complex community structures commonly found in real-world networks.